《Tempus AI（TEM）2024年10-K年度报告「NASDAQ 」（英文版）（267页）.pdf》由会员分享，可在线阅读，更多相关《Tempus AI（TEM）2024年10-K年度报告「NASDAQ 」（英文版）（267页）.pdf（267页珍藏版）》请在三个皮匠报告上搜索。

1、 UNITED STATESSECURITIES AND EXCHANGE COMMISSIONWashington,D.C.20549 FORM 10-K(Mark One)ANNUAL REPORT PURSUANT TO SECTION 13 OR 15(d)OF THE SECURITIES EXCHANGE ACT OF 1934 For the fiscal year ended December 31,2024OR TRANSITION REPORT PURSUANT TO SECTION 13 OR 15(d)OF THE SECURITIES EXCHANGE ACT OF

2、1934 FOR THE TRANSITION PERIOD FROM TO Commission File Number 001-42130 Tempus AI,Inc.(Exact name of Registrant as specified in its Charter)Delaware47-4903308(State or other jurisdiction ofincorporation or organization)(I.R.S.EmployerIdentification No.)600 West Chicago Avenue,Suite 510 Chicago,IL 60

3、654(Address of Principal Executive Offices,Zip Code)Registrants telephone number,including area code:(800)976-5448 Securities registered pursuant to Section 12(b)of the Act:Title of each class TradingSymbol(s)Name of each exchange on which registeredClass A common stock,$0.0001 par value per share T

4、EM The Nasdaq Stock Market LLCSecurities registered pursuant to Section 12(g)of the Act:NoneIndicate by check mark if the Registrant is a well-known seasoned issuer,as defined in Rule 405 of the Securities Act.YES NO Indicate by check mark if the Registrant is not required to file reports pursuant t

5、o Section 13 or 15(d)of the Act.YES NO Indicate by check mark whether the Registrant:(1)has filed all reports required to be filed by Section 13 or 15(d)of the Securities Exchange Act of 1934 during the preceding 12 months(or for such shorter period that the Registrant was required to file such repo

6、rts),and(2)has been subject to such filing requirements for the past 90 days.YES NO Indicate by check mark whether the Registrant has submitted electronically every Interactive Data File required to be submitted pursuant to Rule 405 of Regulation S-T(232.405 of this chapter)during the preceding 12 m

7、onths(or for such shorter period that the Registrant was required to submit such files).YES NO Indicate by check mark whether the registrant is a large accelerated filer,an accelerated filer,a non-accelerated filer,smaller reporting company,or an emerging growth company.See the definitions of“large

8、accelerated filer,”“accelerated filer,”“smaller reporting company,”and“emerging growth company”in Rule 12b-2 of the Exchange Act.Large accelerated filer Accelerated filer Non-accelerated filer Smaller reporting company Emerging growth company If an emerging growth company,indicate by check mark if t

9、he Registrant has elected not to use the extended transition period for complying with any new or revised financial accounting standards provided pursuant to Section 13(a)of the Exchange Act.Indicate by check mark whether the Registrant has filed a report on and attestation to its managements assess

10、ment of the effectiveness of its internal control over financial reporting under Section 404(b)of the Sarbanes-Oxley Act(15 U.S.C.7262(b)by the registered public accounting firm that prepared or issued its audit report.If securities are registered pursuant to Section 12(b)of the Act,indicate by chec

11、k mark whether the financial statements of the Registrant included in the filing reflect the correction of an error to previously issued financial statements.Indicate by check mark whether any of those error corrections are restatements that required a recovery analysis of incentive-based compensati

12、on received by any of the Registrants executive officers during the relevant recovery period pursuant to 240.10D-1(b).Indicate by check mark whether the Registrant is a shell company(as defined in Rule 12b-2 of the Exchange Act).YES NO The aggregate market value of the voting and non-voting common e

13、quity held by non-affiliates of the Registrant,as of the last business day of the Registrants most recently completed second fiscal quarter was approximately$5.4 billion(based on the closing price of the Registrants Class A common stock on the Nasdaq Global Select Market on June 28,2024 of$35.00 per

14、 share).As of February 21,2025,there were 167,168,911 shares of Class A common stock and 5,043,789 shares of Class B common stock,each with a par value of$0.0001 per share,outstanding.DOCUMENTS INCORPORATED BY REFERENCE Portions of the Registrants definitive proxy statement,or the 2025 Proxy Stateme

15、nt,relating to its annual meeting of stockholders to be held in 2025,or the 2025 Annual Meeting,to be filed with the Securities and Exchange Commission,or the SEC,within 120 days after the end of the fiscal year to which this Annual Report on Form 10-K relates,are incorporated herein by reference wh

16、ere indicated.Except with respect to information specifically incorporated by reference in this Annual Report on Form 10-K,such proxy statement is not deemed to be filed a part hereof.iTable of Contents PagePART I Item 1.Business 3Item 1A.Risk Factors 51Item 1B.Unresolved Staff Comments 114Item 1C.C

17、ybersecurity 114Item 2.Properties 116Item 3.Legal Proceedings 116Item 4.Mine Safety Disclosures 117 PART II Item 5.Market for Registrants Common Equity,Related Stockholder Matters and Issuer Purchases of Equity Securities 118Item 6.Reserved 119Item 7.Managements Discussion and Analysis of Financial

18、Condition and Results of Operations 120Item 7A.Quantitative and Qualitative Disclosures About Market Risk 136Item 8.Financial Statements and Supplementary Data 137Item 9.Changes in and Disagreements With Accountants on Accounting and Financial Disclosure 178Item 9A.Controls and Procedures 178Item 9B

19、.Other Information 178Item 9C.Disclosure Regarding Foreign Jurisdictions that Prevent Inspections 179 PART III Item 10.Directors,Executive Officers and Corporate Governance 180Item 11.Executive Compensation 180Item 12.Security Ownership of Certain Beneficial Owners and Management and Related Stockho

20、lder Matters 180Item 13.Certain Relationships and Related Transactions,and Director Independence 180Item 14.Principal Accounting Fees and Services 180 PART IV Item 15.Exhibits,Financial Statement Schedules 181Item 16.Form 10-K Summary 184 1NOTE REGARDING FORWARD-LOOKING STATEMENTS This Annual Report

21、 on Form 10-K,including the sections titled“Business”and“Managements Discussion and Analysis of Financial Condition and Results of Operations,”contains forward-looking statements about us and our industry that involve substantial risks and uncertainties.All statements other than statements of histor

22、ical facts contained in this Annual Report on Form 10-K or incorporated herein,including statements regarding our future results of operations or financial condition,business strategy and plans and objectives of management for future operations,are forward-looking statements.In some cases,you can id

23、entify forward-looking statements because they contain words such as“anticipate,”“believe,”“contemplate,”“continue,”“could,”“estimate,”“expect,”“intend,”“may,”“plan,”“potential,”“predict,”“project,”“should,”“target,”“will”or“would”or the negative of these words or other similar terms or expressions.

24、Within the meaning of Section 27A of the Securities Act of 1933,as amended,or the Securities Act,and Section 21E of the Securities Exchange Act of 1934,as amended,or the Exchange Act,these forward-looking statements contained in this Annual Report on Form 10-K include,but are not limited to,statemen

25、ts concerning the following:the evolving treatment paradigm for cancer,including physicians use of molecular data and targeted oncology therapeutics and the market size for our current and future products;our ability to expand our business beyond oncology into new disease areas;estimates of our addr

26、essable market and our expectations regarding our revenue,expenses,capital requirements and operating results;our ability to develop new products and services,including our goals and strategy regarding development and commercialization of AI Applications;our ability to maintain and grow our datasets

27、,including in new disease areas and geographies;any expectation that the growth of our datasets will improve the quality of our products and services and accelerate their adoption;our ability to capture,aggregate,analyze or otherwise utilize genomic data in new ways and in additional diagnostic moda

28、lities;any expectation that we will continue to commercialize de-identified records and license them to multiple customers;the acceptance of our publications in peer-reviewed journals or of our presentations at scientific and medical conference presentations;the implementation of our business model

29、and strategic plans for our products,technologies and businesses;competitive companies and technologies and our industry;the potential of Intelligent Diagnostics to be disruptive across a broad set of disease areas and the clinical trial process;our ability to manage and grow our business by expandi

30、ng our sales to existing customers or introducing our products to new customers;third-party payer reimbursement and coverage decisions,including our strategy to increase reimbursement;our ability to establish and maintain intellectual property protection for our products or avoid claims of infringem

31、ent;potential effects of evolving and/or extensive government regulation;the timing or likelihood of regulatory filings and approvals;our ability to hire and retain key personnel;our ability to expand internationally,including through our joint venture,SB Tempus Corp.,or SB Tempus,in Japan;our abili

32、ty to successfully acquire businesses,form joint ventures or make investments in companies or technologies,including our ability to realize the expected benefits of strategic plans,including our acquisition of Ambry Genetics Corporation,or Ambry;2our ability to protect and enforce our intellectual p

33、roperty rights,including our trade secret protected proprietary rights in our platform;our ability to service or pay down existing or future debt obligations;our anticipated cash needs and our needs for additional financing;andanticipated trends and challenges in our business and the markets in whic

34、h we operate.You should not rely on forward-looking statements as predictions of future events.We have based the forward-looking statements contained in this Annual Report on Form 10-K primarily on our current expectations and projections about future events and trends that we believe may affect our

35、 business,financial condition and operating results.The outcome of the events described in these forward-looking statements is subject to risks,uncertainties and other factors described in Part I,Item 1A,“Risk Factors”in this Annual Report on Form 10-K.Moreover,we operate in a very competitive and r

36、apidly changing environment.New risks and uncertainties emerge from time to time,and it is not possible for us to predict all risks and uncertainties that could have an impact on the forward-looking statements contained in this Annual Report on Form 10-K.The results,events and circumstances reflecte

37、d in the forward-looking statements may not be achieved or occur,and actual results,events or circumstances could differ materially from those described in the forward-looking statements.In addition,statements that“we believe”and similar statements reflect our beliefs and opinions on the relevant su

38、bject.These statements are based on information available to us as of the date of this Annual Report on Form 10-K.And while we believe that information provides a reasonable basis for these statements,that information may be limited or incomplete.Our statements should not be read to indicate that we

39、 have conducted an exhaustive inquiry into,or review of,all relevant information.These statements are inherently uncertain,and investors are cautioned not to unduly rely on these statements.The forward-looking statements made in this Annual Report on Form 10-K relate only to events as of the date on

40、 which the statements are made.We undertake no obligation to update any forward-looking statements made in this Annual Report on Form 10-K to reflect events or circumstances after the date of this Annual Report on Form 10-K or to reflect new information or the occurrence of unanticipated events,exce

41、pt as required by law.We may not actually achieve the plans,intentions or expectations disclosed in our forward-looking statements,and you should not place undue reliance on our forward-looking statements.Our forward-looking statements do not reflect the potential impact of any future acquisitions,m

42、ergers,dispositions,joint ventures or investments.3PART IItem 1.Business.OverviewWe endeavor to unlock the true power of precision medicine by creating Intelligent Diagnostics through the practical application of artificial intelligence,or AI,in healthcare.Intelligent Diagnostics use AI,including ge

43、nerative AI,to make laboratory tests more accurate,tailored,and personal.We make tests intelligent by connecting laboratory results to a patients own clinical data,thereby personalizing the results.Our novel insight was realizing that all laboratory test results,genomic or otherwise,could be context

44、ualized for a specific patient based upon that patients unique characteristics,and technology could therefore guide therapy selection and treatment decisions to allow each patient to progress on their own unique path.The drugs recommended,the clinical trials explored,the care pathways evaluated,the

45、adverse events consideredall have the potential to be refined and enhanced when test results are connected to a patients personal profile,enabling the right patient to be routed to the right therapy at the right time.To accomplish this,we built the Tempus Platform,which comprises both a technology p

46、latform to free healthcare data from silos and an operating system to make the resulting data useful.Our proprietary technology has allowed us to amass what we consider to be one of the largest libraries of clinical and molecular oncology data in the world.Our goal is to embed AI,including generativ

47、e AI,throughout every aspect of diagnostics to enable physicians and researchers to make personalized,data-driven decisions that improve patient care.The ability to deploy AI in precision medicine at scale has only recently become possible.Advances in cloud computing,imaging technologies,large langu

48、age models and low-cost molecular profiling,along with the digitization of vast amounts of healthcare data,have created a landscape that we believe is finally ripe for AI.However,despite an increase in the availability of healthcare data,physicians and researchers are largely unable today to leverag

49、e this data to improve patient care.The vast majority of healthcare data remains disconnected and lacks harmonization and structure.Traditional diagnostic tests are typically based only on a single data modality,such as a blood-based biomarker or a genomic mutation,and do not connect and integrate o

50、ther forms of relevant clinical data,such as outcomes,or adverse events,or pathology results,which are essential for many clinical decisions.In order to bring AI to healthcare at scale,we believe the foundation of how data flows throughout the ecosystem needs to be rebuilt.We established new data pi

51、pes,going to and from providers,to allow for the free exchange of data between physicians,who interpret data,and diagnostic and life science companies,who provide data,integrating relevant clinical data,such as outcomes,or adverse events,which are essential for many clinical decisions.Without this c

52、apability,we believe that data would continue to accumulate without impacting patient care.To accomplish this,we built both a technology platform to free healthcare data from silos and an operating system to make this data useful,the combination of which we refer to as our Platform.Our Platform conn

53、ects multiple stakeholders within the larger healthcare ecosystem,often in near real time,to assemble and integrate the data we collect,thereby providing an opportunity for physicians to make data-driven decisions in the clinic and for researchers to discover and develop therapeutics.We aim to help

54、physicians find the best therapies for their patients,help pharmaceutical and biotechnology companies make the best drugs possible,and enable patient access to emerging therapies and clinical trials when appropriate.Tempus is a technology company focused on healthcare that straddles two converging w

55、orlds.We strive to combine deep healthcare expertise,providing next-generation diagnostics across multiple disease areas,with leading technology capabilities,harnessing the power of data and analytics to help personalize medicine.We endeavor to unlock the true power of precision medicine by creating

56、 Intelligent Diagnostics through the practical application of artificial intelligence,or AI,in healthcare.Intelligent Diagnostics use AI,including generative AI,to make laboratory tests more accurate,tailored,and personal.Unlike traditional diagnostic labs,we can incorporate unique patient informati

57、on,such as clinical,molecular,and imaging data,with the goal of making our tests more intelligent and our results more insightful.Unlike other technology companies,we are deeply rooted in clinical care delivery as one of the largest sequencers of cancer patients,and patients with other diseases,in t

58、he United States.Straddling both worlds is advantageous as we believe Intelligent Diagnostics represent the future of precision medicine,informing more personalized and data-driven therapy selection and development.We believe their adoption could empower physicians to deliver better care and researc

59、hers to develop more precise therapies,with the potential to save millions of lives.4Our Platform includes proprietary software and dedicated data pipelines that create a network of healthcare institutions through approximately 700 unique data connections,many of which supply us with complex multimo

60、dal data in near real time,across approximately 3,000 healthcare institutions that order our products and services.Healthcare institutions supply us with this data in our capacity as a covered entity(for example,when we provide Next Generation Sequencing,or NGS,services on behalf of a patient),or as

61、 a business associate(for example,when we provide clinical trial matching services or data de-identification and structuring services).In addition to the data we receive in these capacities,we currently have a limited number of paid license agreements through which we acquire de-identified data dire

62、ctly from healthcare associations or institutions,and in certain circumstances we cover the actual direct costs associated with the technical integrations needed to create a data connection.We then integrate this data into a unified multimodal database through which we offer numerous analytical and

63、decision support capabilities to our customers.We establish dedicated and integrated data connections with healthcare institutions to enhance the information we provide in our clinical reports,to increase the effectiveness of our clinical trial matching services,and to enable our AI Applications pro

64、duct line,which we believe has the ability to transform healthcare.We have developed multiple productseach based on our Platformthat have allowed us to invest in structuring and harmonizing multimodal data,which is a necessary precursor for deploying AI at scale.Our products are organized under thre

65、e product lines,Genomics,Data,and AI Applications or Algos.Each product line is designed to enable and enhance the others,thereby creating network effects in each of the markets in which we operate.Our business model allows pharmaceutical and biotechnology companies to unlock value from the data we

66、collect,and allows us to monetize a de-identified copy of that data,in different ways across our different product lines.We believe these network effects provide a unique advantage to our business as the compounding value of each data record in our database serves to enhance our competitive moat.The

67、 more data we collect,the smarter our tests become,the more applications we launch,the more physicians join our network,further growing our database,making our tests more precise for clinicians and our database more valuable for researchers.Our Genomics product line leverages our laboratories to pro

68、vide NGS diagnostics,PCR profiling,and other anatomic and molecular pathology testing to healthcare providers,life sciences companies,researchers,and other third parties.However,unlike other laboratory diagnostic testing providers,many of our tests are connected to clinical data in some manner,which

69、 allows our suite of tests to be self-learning and become more accurate with each new test that we run.Our Data and Services product line facilitates drug discovery and development for life sciences companies through two primary products,Insights and Trials.Through our Insights product,we license de

70、-identified libraries of linked clinical,molecular,and imaging data and provide a suite of analytic and cloud-and-compute tools to pharmaceutical and biotechnology companies.Our second product within our Data and Services product line,Trials,leverages the broad network of physicians we work with in

71、oncology to provide clinical trial support for pharmaceutical companies that are looking to reach hard-to-find and underserved patient populations.Our third product line,AI Applications,is focused on developing and providing diagnostics that are algorithmic in nature,implementing new software as a m

72、edical device,and building and deploying clinical decision support tools.The primary product of AI Applications is currently“Next,”an AI platform that leverages machine learning to apply an“intelligent layer”onto routinely generated data to proactively identify and minimize care gaps for oncology an

73、d cardiology patients.As this product gains adoption,we intend to leverage large language models,generative AI algorithms,and our vast database of de-identified data to develop algorithmic diagnostics designed to identify these patients earlier in their disease progression,when treatments are most e

74、ffective.5Industry Background The Limitations of Employing Technology,Data,and AI in Healthcare and Precision MedicineTechnology has had a significant impact on almost every sector of our global economy.From the way we shop online,access information on the internet,or use GPS to navigate the world.W

75、e benefit from,and depend on,technology,data,and the vast computational and connective ecosystem that surrounds us.Yet healthcare has seemingly lagged other industries in embracing the power of technology and leveraging the ensuing computational revolution.We believe this is changing.Recent technolo

76、gical advancements have facilitated the deployment of modern computational methods,such as AI and machine learning,to improve healthcare.Breakthroughs in cloud computing,imaging technologies,large language models,and low-cost molecular profiling have made it easier and more cost effective to digitiz

77、e,structure,harmonize,and store healthcare data,and analyze the resulting datasets at an unprecedented rate.These developments are expediting the adoption of AI,which we believe will impact all aspects of healthcare,from clinical diagnostic testing to the discovery and development of therapeutics,to

78、 healthcare delivery more broadly.Despite the accumulation of healthcare data,we believe the healthcare system still lacks the integrated networks and modern analytical tools necessary to facilitate data-driven care at scale.The vast majority of healthcare data created today remains locked in silos

79、and lacks harmonization due to decentralized institutions using non-standardized methods for collecting data,in addition to a large percentage of the data being in unstructured formats like free text(such as physician progress notes)and non-digitized images(such as pathology slides).Clinical outcome

80、s data,to the extent it even exists,often remains disconnected from diagnostic data,and traditional laboratory tests provide results that are often based only on a single data modality that lack patient context.In addition,clinical and research decisions are too often made based on small sample size

81、s of historic data.In order to bring AI to healthcare at scale,we began by rebuilding the foundation of how data flows in and out of healthcare institutions,which we refer to as the Tempus Platform.We have established data pipes,going to and from providers,which allow for the free exchange of data b

82、etween physicians,who interpret data,and diagnostic and therapeutic companies,who provide data.Harnessing the power of this data at scale required a Platform that could break down data silos,collect vast amounts of multimodal data,structure and harmonize it,and deploy AI to make it useful for physic

83、ians and researchers to make data-driven decisions in the clinic or at the lab bench,thereby advancing precision medicine.Our access to broad and diverse data serves as the basis for our ability to train generative AI models,and we believe our relationships with healthcare institutions provide us wi

84、th proprietary data to deliver on the promise of AI in healthcare.Without this Platform,we believe the data would continue to pile up at an increasing rate without improving patient care.We have built a version of this Platform and are now deploying it at scale in oncology in the United States,with

85、other disease areas following.Importance of Multimodal Healthcare DataTechnology is enabling the healthcare industry to collect data at an unprecedented scale,yet most datasets continue to be fractured or narrowly focused by disease type or data modality;almost none are comprehensive enough to provi

86、de a full picture of the patient and their clinically relevant characteristics.We set out to solve that problem by building a platform that collects broad datasets in near real time and at scale.Our Platform is differentiated in several ways.First,we collect data from multiple diagnostic modalities,

87、including NGS,anatomic pathology slides,radiology images,and other laboratory tests.Second,the data we collect is often connected to EHR data,such as key phenotypic characteristics,therapeutic data,and clinical outcome and response data.Third,our Platform is multi-disease,spanning oncology,neurology

88、,cardiology,and infectious disease.Our Platform is purpose built to deploy AI at scale,using multimodal datasets,across disease areas.We believe these differentiators have the potential to transform healthcare.A New Industry:Intelligent Diagnostics to Advance Precision MedicineWhile AI has the poten

89、tial to broadly impact healthcare,we believe it will transform diagnostics first.Diagnostics,broadly defined,is the process of determining by examination or assessment the nature and circumstance of disease.Physicians use diagnostics all the time;they order blood tests,biopsies,scans,genomic tests,a

90、nd others.Physicians rely on diagnostic results to make the vast majority of their treatment decisions.Researchers rely on diagnostic tests to better understand disease and make better decisions throughout their discovery processes.6The ability to leverage generative AI on top of large,harmonized,mu

91、ltimodal datasets provides the opportunity to make diagnostic tests more personalized,and therefore more intelligent.Intelligent Diagnostics incorporate an individual patients longitudinal phenotypic,morphologic,and molecular data,including outcome data from the patients EHR,to give laboratory test

92、results clinical context.In doing so,Intelligent Diagnostics can leverage generative AI to make laboratory tests more accurate,tailored,and personal.The test result itself is designed to be specific to each patient and their own unique patient journey.The result is also informed by our large dataset

93、 that enables association of clinical outcomes and therapeutic response for patients who are similar to the patient being treated.The process for making a diagnostic“intelligent”improves upon the process for performing genomic testing,by leveraging technology and data to add clinical context and the

94、rapeutic insights.An Intelligent Diagnostic requires the following:(i)perform a laboratory test or ingest results from a laboratory test;(ii)review the test results on a stand-alone basis;(iii)combine the stand-alone results with other forms of relevant clinical data from that patients medical recor

95、ds;(iv)contextualize or reconfigure the stand-alone laboratory results to the extent necessary with the insight derived from that patients clinical history;(v)include the outcome and response data of patients who are similarly situated to the patient for whom the test was ordered;and(vi)use generati

96、ve AI to derive analytical and clinically relevant insights and provide those to the physician and patient.See below for an illustration comparing an Intelligent Diagnostic to a standard genomic test:We believe the adoption and deployment of Intelligent Diagnostics will have a substantial impact on

97、patient care.In oncology,for example,Intelligent Diagnostics have the potential to eventually incorporate insights using data from molecular and anatomic pathology,bioinformatics,genomic variant analysis,inherited cancer risk,computational biology,drug label data,noted adverse events,clinical trial

98、data,research publications,investigational studies,care pathways,real world evidentiary studies,and phenotypic and morphologic data.We already have the ability to incorporate many of these data elements today.The consequence of incorporating multimodal data is to make precision medicine“personalized

99、”as opposed to“targeted.”A targeted diagnostic test might find a specific condition or characteristic of a patient that is relevant to a particular therapy.For example,in cancer,a targeted diagnostic test may identify a genomic biomarker that could inform therapy selection,such as identifying a HER2

100、 amplification that would allow a HER2 inhibitor to be prescribed to a breast cancer patient.The standard test to determine whether a HER2 amplification is present(other than at Tempus)is typically not designed to assess factors such as whether the patient is male or female,old or young,or has diabe

101、tes or a heart condition.Nor does the standard test consider the medication the patient has taken or is currently taking,or the adverse events the patient has experienced.7 An Intelligent Diagnostic test,by contrast,might recommend specific therapies based not just on a singular characteristic,but o

102、n the comprehensive profile of the patient who will receive the proposed therapy.For example,an Intelligent Diagnostic might highlight that the breast cancer patient should consider immunotherapy before taking the HER2 inhibitor,or might highlight a series of adverse events the physician should be a

103、ware of based on other phenotypic characteristics for that patient,such as if the patient had a heart condition and therefore an elevated risk of a cardiac adverse event from taking the HER2 inhibitor.By linking multimodal data regarding both the disease,such as cancer or diabetes,and the host,our t

104、ests can provide a more comprehensive and holistic view of the patient and reconfigure results based in part on the clinical data we collect and the aggregate information in our database.Intelligent Diagnostics also have the potential to disrupt the clinical trial process.Today new therapies are typ

105、ically approved based on randomized clinical trials that apply to broad populations and demonstrate incremental improvements over the existing standard of care.The current process suffers from several inherent flaws.First,clinical trials are generally expensive and slow to complete.Second,if and whe

106、n therapeutics are approved,they can have less of an impact on the larger population than the trial population,given an inherent bias on who has access to academic medical centers and emerging studies.Third,many new therapies are only effective on a subset of patients that enter clinical trials.We b

107、elieve Intelligent Diagnostics,AI,and technology broadly can help solve these problems.We believe our ability to contextualize test results to individual patients,to incorporate real world evidence at scale,to identify patterns across similarly situated patients,will help physicians make better,data

108、-driven decisions which drug to prescribe,which trial to consider,and so on.The Tempus PlatformTempus set out to build proprietary technology to implement Intelligent Diagnostics and to facilitate access to,and use of,the resulting datasets.The Tempus Platform connects multiple stakeholders within t

109、he larger healthcare ecosystem and provides both the technical infrastructure for what we consider to be one of the worlds largest libraries of matched clinical and molecular data,and an operating system to make that information useful.Our Platform is end-to-end and vertically integrated.It allows u

110、s to ingest data from providers,perform diagnostic testing upon request,generate results leveraging our multimodal database,and provide clinical context for a specific patient.Below is a graphic illustrating our Platforms core functionality.8We believe our AI-enabled Platform can provide unique valu

111、e whenever two conditions exist:a heterogeneous diseased population and a variety of therapeutics or therapeutic pathways,which are often prescribed based on trial and error.For example,in oncology,there is a diverse population diagnosed with cancer,and each subtype has different characteristics.The

112、 combination of unique patient characteristics and different cancer subtypes results in a variety of phenotypic attributes(old,young,male,female,black,white,etc.).In addition,there are hundreds of possible therapeutic paths to consider in cancer(surgery,radiotherapy,chemotherapy,targeted therapy,imm

113、unotherapy,etc.).These conditions create an ideal backdrop for the benefits of big data and AI.The same is true in neuropsychiatry.A heterogeneous population suffers from numerous neurological disorder subtypes,such as depression,anxiety,bipolar disorder,and other psychiatric conditions.Like oncolog

114、y,there is a diverse patient population and a number of prescribed antidepressants,often based on trial and error.Further,the complexity of oncology,neuropsychiatry,and many other major causes of morbidity necessitate a multimodal data approach,as any single modality(e.g.,DNA-only)is unlikely to pro

115、vide enough information to differentiate meaningful patient subgroups.We believe technology and AI should facilitate data associations and substantially reduce the guesswork associated with which drug to prescribe,in what amount,and in which order.Facilitated by our relationships with many leading h

116、ospitals across the healthcare system in the United States,we believe we are well positioned to introduce precision medicine at scale across multiple disease categories and drive adoption of our Platform and novel AI solutions.We are leveraging our ability to collect,structure and harmonize data,and

117、 deploy AI on large datasets to facilitate precision medicine broadly.We initially deployed our Platform in oncology,expanded substantially within oncology,and recently extended into neuropsychiatry,radiology,and cardiology.Below is a timeline of our Platforms evolution,both within oncology and into

118、 different disease categories:Core Elements of our Platform The Tempus Platform combines multiple elements into a vertically integrated infrastructure that enables us to ingest data from providers,structure and harmonize the data into a common database,provide laboratory diagnostic testing,and deliv

119、er personalized results that provide clinical context by leveraging our data.We offer closed-loop,full-stack,bi-directional integrations between a clinicians desktop and our laboratory diagnostic capabilities,analytics platform,and repository of multimodal data.Our scaled,interconnected provider net

120、work covers more than 50%of U.S.oncologists and provides us with broad data rights,including the rights to longitudinally updated data from time to time.The combination of our Platform and vast provider network yields a powerful flywheel that continues to become more accurate and precise as more pat

121、ients are added,thereby compounding the network effects of our offering.We believe each of these elements is difficult for competitors to replicate,and together represent a significant competitive advantage.The following diagram represents the different elements of our Platform.9 Ingestion and Gener

122、ation of Data We ingest healthcare data in near real time and at scale,including molecular,clinical,and imaging data.Between our sequencing and data collection efforts,we are connected in some way to more than 50%of all oncologists practicing in the United States,along with a growing number of patie

123、nts in neuropsychiatry,cardiology,and infectious disease.Our methods for collecting and creating data include the following:Ingesting data through our relationships and partnerships with healthcare providers.We have developed proprietary tools to establish approximately 700 direct data connections,a

124、cross approximately 3,000 hospitals,many of which are bi-directional.We have established relationships with hundreds of provider networks,including more than 65%of all academic medical centers in the United States.To obtain data from these sources,we use a variety of near real-time connections(e.g.,

125、HL7,FHIR)and batch data exchanges.Healthcare institutions supply us with this data in our capacity as a covered entity(for example,when we provide NGS services on behalf of a patient),or as a business associate(for example,when we provide clinical trial matching services or data de-identification an

126、d structuring services).We ingest and structure data using optical character recognition,or OCR,natural language processing,or NLP,and proprietary workflow tools along with manual data curation.Our proprietary tools connect to a providers EHR system,data warehouse,or third-party data provider to pul

127、l out relevant structured and unstructured data that the provider has agreed to provide to Tempus,including longitudinal follow-up data to the extent the provider has made such data available.To facilitate these data-sharing relationships,we have developed software products and services that align t

128、o our customers interests by helping providers use our software tools to improve patient care.In certain circumstances,we cover the actual direct costs associated with the technical integrations needed to create a data connection.We cover these costs to help facilitate providers contribution of data

129、 and their corresponding use of our products,which then makes our tests more intelligent and helps them to facilitate the delivery of better care.We generally retain the rights we acquire in de-identified data even if our contractual obligations expire or are terminated.Relationships with industry a

130、ssociations.In addition to healthcare providers,we work with numerous industry associations in the United States,such as ASCO.Under our collaboration with ASCO,we structure and distribute the oncology data ASCO collects as part of CancerLinq,which is their oncology data effort.We work with other lar

131、ge associations such as ONCare Alliance,LLC(surviving entity after merger of the National Cancer Care Alliance and the Quality Cancer Care Alliance),and have agreements in place with large integrated community practices.While our 10relationships in oncology are widespread,we are making inroads in ot

132、her disease areas.For example,we are working with a large hospital network to train algorithmic models based on a de-identified subset of approximately 3.5 million electrocardiograms,or ECGs,across more than 800,000 patients,with decades of longitudinal clinical data,including outcome and response d

133、ata.We also have agreements with numerous other institutions through both our sequencing and data efforts to collect and structure multimodal infectious disease data,and have entered into a variety of partnerships and collaborations across neuropsychiatry,diabetes,and cardiology giving us access to

134、additional clinical data.Laboratory diagnostics.In addition to our dedicated data pipelines,we generate data for our Platform from our four high-throughput diagnostic testing labs in Chicago,Atlanta,Raleigh,and Aliso Viejo.Our labs offer a range of anatomical and molecular NGS tests,including a broa

135、d portfolio of solid tumor and liquid biopsy cancer tests.Our laboratory offerings enable us to populate our database with connected and comprehensive molecular,clinical,and morphologic data that has been de-identified.We also make available an unrestricted copy of the raw files containing the rich

136、data we generate in the laboratory,along with any clinical data we curate,to the providers who order our tests,to further enable their own research efforts.In February 2025,we acquired Ambry Genetics Corporation,or Ambry,a leader in hereditary cancer screening and the supplier of our germline sequen

137、cing(Tempus|xG)for hereditary cancer risk.With the acquisition of Ambry,we can leverage its vast amounts of data to augment our current data offerings.Further,Ambrys offerings span multiple disease areas,enabling us to expand beyond oncology into new categories,such as pediatrics,rare disease,cardio

138、logy,reproductive health and immunology.Additionally,Ambrys significant laboratory capabilities on the west coast will help increase our overall footprint in the country.We ingest and generate a variety of different types of data from different sources.The following represents selected data modaliti

139、es that we collect and aggregate into our database.Proprietary Data ProcessingOnce data is ingested,we deploy proprietary clinical data abstraction tools,including natural language processing,optical character recognition,and our abstraction software,to structure,harmonize,and de-identify the data w

140、e collect.We have developed various software tools,including algorithmic agents that leverage large language models,to organize millions of records into a common format that spans a variety of data types.For example,we organize clinical data from unstructured documents and structured EHR fields,and

141、typically digitize whole-slide pathology images as part of our clinical workflow.We then combine this data with the molecular data that we generate in our labs or process from third parties,giving us a more comprehensive profile of patients.Unstructured data housed in physician notes and other docum

142、ents is processed using OCR and NLP,mapped to Tempus Medical Ontology,and routed to data abstractors for further curation and quality control.Typically we receive identified data,either in our capacity as a covered entity under the Health Insurance 11Portability and Accountability Act,or HIPAA,or to

143、 the extent we have a business associate agreement with the provider.Following abstraction and structuring,we de-identify data and only retain the resulting de-identified dataset,other than through our obligations to retain selected identified data as a covered entity providing laboratory tests to c

144、linicians.Many clinicians who order Tempus tests clinically are also involved in research related activities.By making this organized and structured data available to the clinicians(along with raw files associated with the testing we perform)we serve,those clinicians can use the data to further thei

145、r own research efforts to help patients.Our Proprietary Multimodal DatabaseWe believe most healthcare databases lack real-time functionality,depth among data types,and the scale of matched clinical and molecular records needed to meaningfully improve therapeutic research and development.Tempus is at

146、tempting to solve this problem by democratizing the use of near-real time molecular,clinical,and imaging data by embedding our solution into the clinical care of patients.As our testing volume has grown,and as our dedicated data pipelines have expanded,the size of our database has increased exponent

147、ially.Since we launched our Platform in 2016,Tempus has amassed over 900 million documents,across more than 7.3 million de-identified patient records,including over 1.1 billion pages of rich clinical text that we use to train our large language models.The database also includes over 1,400,000 record

148、s with imaging data,more than 1,300,000 with matched clinical records linked with genomic information,and more than 260,000 with full transcriptomic profiles.Within oncology specifically,we believe this represents one of the largest and most comprehensive molecular libraries of cancer patients in th

149、e world.The breadth of our database,the quality and diversity of our data,as well as its regularly updating nature,allow us to offer a variety of AI-enabled solutions to the market.We believe our unique data set enables us to bring the benefits of generative AI and large language models to healthcar

150、e,as our curated,multimodal database can be used as a proprietary training set to build a variety of AI based applications,which we intend to deploy through our existing network and distribution platform.We also retain the rights to broadly commercialize de-identified data.As the amount of data in o

151、ur cloud environment continues to grow from its current size of more than 240 petabytes,we believe new AI applications and opportunities will emerge that are only possible with scale,driving innovations in patient treatment that were previously unattainable.The following diagram represents the growt

152、h of our database over time.12Another valuable attribute of our dataset is the number of different data modalities represented.We believe multimodal data is a necessary predicate to successfully build and deploy AI-based applications given the complexity of disease and the various attributes across

153、different forms of data(e.g.,text,images,molecules,etc.).As of December 31,2024,our database included the following types of data,among others:Footnote:Our clinical data typically includes the following information to the extent provided and abstracted by Tempus:unique identifier;age;sex;race/ethnic

154、ity;histology;stage of disease;sample type(primary vs.metastatic);anatomical site of sample and method of procurement;cancer treatment history,including therapies administered;timing of relapse and timing of treatments,including cancer-related treatments and surgery;genomic profiling results(e.g.,in

155、ternal,external providers);tumor response;progression free survival;RECIST or equivalent;ECOG/Karnofsky scores,or equivalent;and adverse events.Proprietary Software Tools and SolutionsWe have developed numerous software tools and applications to help make our services accessible to multiple constitu

156、encies within the healthcare ecosystem and support our various product lines.We believe this system architecture,which employs AI techniques such as neural networks,deep learning,large language models,and other statistical learning techniques to generate patient-specific insights.We are able to not

157、only train and validate some of these AI models for research use,but we can also develop them into clinical-grade algorithmic tests,or Algos,and deploy them clinically as part of routine care.As our data advantage and system architecture continue to improve,we believe our existing Intelligent Diagno

158、stics will gain further adoption thereby accelerating our ability to deploy technologies,including AI Applications,in the clinical setting.We are both a healthcare company and a technology company,which we believe allows us to more quickly and effectively develop and deploy AI techniques within our

159、proprietary software systems.To do so,we rely on employees with expertise spanning multiple disciplines,including those with PhDs and other advanced degrees in the scientific fields of machine learning,data science and computational biology,as well as Medical Doctors practicing disciplines such as p

160、athology and oncology.In addition to our diverse employee base,we are able to train AI models using our proprietary and expansive multimodal,de-identified dataset.We leverage our varied expertise and extensive resources to continuously monitor and review the statistical performance of the models use

161、d across our Platform to ensure performance and prevent degradation.13 We describe below some of the core software applications that form part of our Platform,including examples where we have developed and deployed AI techniques.External Facing ApplicationsWe have two primary software applications t

162、hat serve as interfaces for different markets and allow our customers to interact with our Platform.Hub is our clinical application for physicians and other healthcare providers and is used primarily in our Genomics product line as an end-to-end application for healthcare providers who use our NGS t

163、ests.Lens is our application for life sciences customers and other healthcare researchers,launched in May 2021.Lens is aligned with Insights,one of our products within Data,and allows users to identify,license,and ultimately analyze cohorts of data for research purposes.We typically enable our custo

164、mers to access free or charge certain software applications(like Hub)and certain features of other applications(like Lens).However,in some cases we may charge for access to Lens when a customer is interested in some form of customization or access to Lens full suite of capabilities.HubHub can be acc

165、essed on the web or through our mobile applications.Hub enables physicians and other providers to interact with our Platform,place orders for our laboratory tests,track them through the sequencing process,view results,and develop treatment plans using the other information Tempus makes available.Hub

166、 streamlines and automates what previously required a significant investment of both time and resources for those ordering and delivering genomic reports.A physicians experience,through Hub,typically begins with our online ordering feature,which presents providers with Tempus various test options an

167、d guides users through the ordering process.Once Tempus has processed an order and sequenced a specimen,Hub synthesizes information across our various tests,orders,and patients,and presents the information in a consumer-friendly interface.For example,Order Summary synthesizes information from variou

168、s clinical orders,test results,and other information relevant to a patients course of treatment.A typical patient might have multiple sequencing events over time.Hub visually presents all of a patients results side-by-side,so a treating physician can comprehensively view how a patients disease has c

169、hanged over time,including in response to therapy.Hub also provides care teams a robust set of search and filtering tools so they can navigate our Platform.Physicians can use Hub to identify similarly situated patients or patient sub-groups,including by specific molecular alteration.Physicians can a

170、lso export and download the resulting dataset for further analysis.Hub offers additional functionality that goes beyond ordering and presenting clinical results.Our clinical trial system,for example,handles the complexities of matching patients to clinical trials,by synthesizing clinical and molecul

171、ar data matched against inclusion and exclusion criteria for the trial.It even allows physicians to activate their point of care as a clinical trial site,if approved by the trial sponsor,in order to easily enroll patients who would otherwise not have access to experimental therapies.The proprietary

172、features within Hub put powerful analytics in the hands of physicians,allowing them to pursue research opportunities using accessible molecular data,and explore immune insights such as HLA type,immune infiltrates and neoantigens.Finally,Time on Therapy provides physicians a view into the Tempus Prec

173、ision Medicine Library,which includes the treatment paths of patients within our de-identified database who display similar molecular or phenotypic profiles to their own patients.These tools enable new patients to potentially benefit from the experience of those that came before.One example of an AI

174、 model whose results are available within Hub,and which illustrates a typical development and validation process for our AI models,is our tumor origin,or TO,algorithm.Our TO algorithm predicts the site of origin for cancer patients whose primary tumor site is unknown using machine learning models tr

175、ained on tumor RNA expression results from our de-identified multimodal database.We began developing our TO algorithm in 2019,and it was first deployed in a clinical setting in 2021.We developed and trained the TO algorithm,like other machine learning models,by adopting current best practices for AI

176、 model development.For example,in developing the TO algorithm,we explored distinct model architectures(logistic regression,random forests and neural networks)and feature selection methods,and we utilized multiple cross validation techniques using both our own and independent third-party datasets.Aft

177、er its launch,we continue to monitor the performance of the TO algorithm by using advanced statistical methods to detect potential model drift or degradation over time.Each TO prediction is reviewed by our board certified pathologists for consistency with underlying data,and the distribution of expe

178、cted cases is reviewed and assessed against the expected distribution of diagnoses.14 We include below some illustrations depicting some of Hubs capabilities:15 16 17 LensLens is our software application for life sciences and advanced precision research.We designed Lens to expose our multimodal,de-i

179、dentified dataset to two main constituencies:(i)clinicians interested in exploring data related both to their own patients and to similarly situated patients from the broader Tempus dataset,and(ii)pharmaceutical and biotechnology clients that are focused on drug discovery and development and want to

180、 explore our dataset and/or supplement their own analytics with our tools and data.For clinicians,Lens helps users filter our multimodal database to identify groups of patients that meet their research requirements.It allows browsing,segmenting,selecting,and analyzing cohorts of patients using a var

181、iety of clinical,molecular,and demographic characteristics.We generally make these aspects of Lens available to our customers without charge because such access helps our customers identify data cohorts of interest and facilitates data licensing opportunities.In addition to this basic functionality,

182、Lens allows advanced computational users to perform robust analytics using our cloud-and-compute infrastructure and modeling tool set.We launched certain of these advanced features in May 2021,one of which is called Notebooks,a proprietary tool that allows users to run their own AI models within our

183、 cloud-and-compute environment,taking advantage of fast and streamlined access to our data and computational infrastructure,and saving researchers time and money.Over time,we intend to enter into separate subscription agreements,and charge separately,for expanded access to Lens and the increased fun

184、ctionality we intend to provide to our users.We believe that as Lens evolves,it has the potential to redefine life sciences research as investigators can both use our tools for their computational needs and instantly download the data they need for their analysis.We are not aware of any other applic

185、ation in oncology,or any other major disease area,that allows researchers to build large multimodal cohorts,utilize advanced analytics capabilities to explore the data,and download data for deeper analysis in near real time.18We include below some illustrations depicting some of Lens current capabil

186、ities,which we expect to continue to expand and enhance over time:19 Other Software ApplicationsOur software applications extend beyond the oncology space.In the neuropsychiatry space,for example,we have built a series of proprietary and customized applications that are oriented around depression an

187、d other related psychiatric conditions.In addition,we licensed a customized software tool,which we call TempusPRO,that helps track patient reported outcomes,which we integrate into Hub.Patients use the mobile application to complete regular and systematic check-ins,while providers use the tool to vi

188、ew clinical reports and review the patient reported information.We have developed this application to empower providers to make data-driven,personalized treatment decisions,as well as collect outcome measurements on a regular,longitudinal basis in an effort to build one of the largest real-world mul

189、timodal datasets in psychiatry.20Our Three Product LinesOur products are organized under three product lines,with each product line designed to enable and enhance the others,thereby creating network effects in the markets in which we operate.Our Genomics product line provides a broad range of diagno

190、stic testing services to healthcare providers.Our Data and Services product line monetizes de-identified data that we collect and facilitates enrollment in clinical trials,and which at scale has allowed us to provide a series of data related services to our life sciences customers,such as clinical t

191、rial matching.Our AI Applications product line leverages our database to provide diagnostics entirely driven by data,which helps route patients to the optimal therapy and advance research more broadly.Our three product lines and their corresponding product offerings are illustrated in the diagram be

192、low:We believe the interrelated nature of our three product lines is unique.Our business model allows our clients to unlock value from our data,and allows us to monetize that data(in de-identified format),in different ways across our different product lines.We believe these network effects and the c

193、ompounding impact on the value of each data record in our database enhance our competitive advantage.The more data we collect,the smarter our tests become,the more applications we can launch,the more physicians join our network,further growing our database,making our tests smarter for clinicians and

194、 our database more valuable for researchers.GenomicsWe launched our Genomics product line to provide a comprehensive suite of Intelligent Diagnostics to healthcare providers,and to generate a steady stream of molecular data to help fuel growth in our Data and AI Applications product lines.As we run

195、more tests through our laboratories,and as those tests are linked to patient records and clinical outcomes,we 21grow our data assets and leverage them across our other product lines.We operate three laboratories that provide NGS diagnostics,PCR profiling,and other anatomic and molecular pathology te

196、sts.We have broad capabilities across genomic,transcriptomic,proteomic,microbiomic,epigenetic,and methylation-based assays,and our laboratory infrastructure allows us to operate as a high-quality,low-cost NGS provider broadly serving the market.However,unlike other laboratory diagnostic testing prov

197、iders,many of our tests are connected to clinical data,in some manner,which allows our suite of tests to be self-learning,becoming more accurate and precise with each new test that we run.Furthermore,rather than providing a result based on a single data modality,such as a DNA mutation,our Platform l

198、everages data from other modalities and other patients in an effort to be more comprehensive.We are generally paid for our Genomics services by billing insurance companies,or patients directly,who reimburse us for the tests we run,or by billing providers or pharmaceutical companies directly.The foll

199、owing diagram represents a summary of our test offerings as of December 31,2024:Our Oncology TestsOur Platforms first application was in oncology,where we have built a versatile portfolio of cancer tests spanning solid tumors and hematologic malignancies,germline and somatic variants,and tissue and

200、liquid biopsies.Since our inception,our approach to precision oncology has been to provide comprehensive genomic profiling through NGS that enables us to both generate clinically relevant insights that may not be possible with narrower testing approaches,and contribute high-quality molecular informa

201、tion back to providers and to our database.We offer large-panel solid tumor and hematologic testing through multiple assays,with our core clinical assay(xT and xR)offering large panel DNA,RNA full transcriptome,and incidental germline findings through normal blood or saliva analyses.Our current offe

202、rings also include liquid biopsy(xF),whole exome(xE),and hereditary cancer risk(xG).With our acquisition of Ambry in February 2025,we believe we have an opportunity to further expand and enhance our inherited risk screening capabilities for cancer patients.We are also currently validating a treatmen

203、t response monitoring assay.Our oncology tests are differentiated not only because of their breadth,but also because in many cases they are connected to clinical data,which allows us to account for the drugs the patient took historically,how they responded,and for which clinical trials they are actu

204、ally eligible.We endeavor to not recommend drugs for which a patient has been previously prescribed in a prior line of therapy and failed,and not recommend clinical trials they are not eligible to participate in,based on the inclusion or exclusion criteria of the trial.22 The following table lists o

205、ur current oncology test offerings:Lab Tests Launch YearDescriptionOncology tests Tempus|xT2017Designed to detect actionable oncologic targets by sequencing tumor tissue samplesTypically associated with incidental germline testing for matched normal saliva or blood samples,when available Fourth gene

206、ration test that covers648 genes at 500 x coverage spanning approximately 3.6 Mb of genomic space Includes full TCR,BCR,and HLA typing for immuno-oncology,or IO,signatures Detects TMB,MSI,and fusions The test has an approximately 10-day quoted turnaround time.In our analytical validation,we demonstr

207、ated sensitivities 98%for SNVs,92%for rearrangements/fusions,92%for CNVs and indels,and 99.9%for MSI.Premarket approval(PMA)obtained from the FDA in April 2023 Awarded Advanced Diagnostic Laboratory Test(ADLT)statusTempus|xE2018A whole exome cancer assay designed to identify actionable oncologic var

208、iants as well as neoantigens across the exome from tissue samples,thus enabling IO applications Run at 150-250 x media coverage for approximately 650 of the most significant onco-driving mutations and 150-200 x median coverage for more than 19,000genes on the panel Detects TMB,MSI,and fusionsTempus|

209、xF2018Next-generation liquid biopsy assay covering 105 genes at approximately 20,000 x coverage from peripheral blood samples for solid tumors Typically used for oncogenic and resistance mutations that can be detected in cell free DNA,or cfDNA,from a peripheral blood draw In our analytical validatio

210、n,for 0.5%VAF and 30ng of DNA,we demonstrated 99.9%sensitivity for SNVs,98.8%for indels,99.9%for CNVs,and 97.4%for rearrangements and fusions.xF also demonstrated 100%sensitivity concordance with Roche AVENIO ctDNA Expanded Kit for indels,CNVs,and rearrangements.We also demonstrated 99.9%specificity

211、 for SNVs,indels,and fusions,and 96.2%specificity for CNVsTempus|xG2021The xF+version is a 523 gene panel that includes bTMB,MSI,additional fusions and CNVs52 gene inherited cancer germline panel run off whole exome platform at 75x depth of coverage Tests hereditary predisposition across common and

212、well-described cancer syndromes such as breast,ovarian,prostate cancer(BRCA1,BRCA2),pancreatic cancer(CDKN2A,PALB2),colorectal cancer(APC,BMPR1A),and Lynch Syndrome(MLH1,MSH2,MSH6,PMS2,EPCAM)Typically used in patients with a personal and/or family history suggestive of hereditary predisposition to c

213、ancer and can guide future diagnostic decisions The xG+version is an 88 gene panel covering genes associated with both 23Lab Tests Launch YearDescription common and rare hereditary cancersTempus|xR2023Full transcriptomic profiling assay for solid tumors and hematologic malignancies at 50 million pai

214、red end reads,offered as a separate test as of January 2023(previous paired with xT and xE)Reports clinically relevant fusions for more than 100 targeted genes,as well as altered splicing events for MET exon 14 and EGFRvIII,in an unbiased and comprehensive manner 43.4%of patients were matched to a t

215、argeted therapy when DNA seq,RNA seq,and immune biomarker assessment were combined,compared to 29.6%of patients who had a therapy match using DNA seq alone Among patients with identified fusions,29%more patients were identified with a unique clinically actionable fusion that could be matched to a ta

216、rgeted therapy when RNA seq was incorporated,compared to DNA seq alone The test has an approximately 10-day quoted turnaround time We launched xM,a high coverage methylation sequencing assay for monitoring for cancer recurrence and minimal residual disease on June 1,2024,initially covering colorecta

217、l cancer with the potential to expand into additional indications.In November 2023,we entered into a Commercialization and Reference Laboratory Agreement with Personalis,Inc.,or Personalis,pursuant to which we began marketing Personalis Personal Dx test in the United States initially in non-small ce

218、ll lung cancer and breast cancer,as well as IO treatment response monitoring.Personalis will conduct additional development activities to further analytically validate the test in other indications.Personalis will perform tests ordered by patients through us and will bill such patients or payers.We

219、believe incorporating clinical data in our diagnostic tests has widespread benefits.For example,combining clinical and molecular data resulted in improved therapy matching for patients in a study that we conducted,the results of which were published in Nature Bio in September 2019.In that study,usin

220、g our sequencing results and matched clinical data from 500 patient samples across a range of tumor types,we observed that 96%of patients could be matched to at least one clinical trial.Approximately 77%of patients were matched to at least one clinical trial based on a gene variant.Of the patients w

221、ho were not matched to a biomarker-based clinical trial,19.4%were matched to at least one disease-based clinical trial from clinical data alone.24The results of the Nature Bio study indicated that paired tumor-normal DNA-seq and RNA profiling of patient cancer biopsies yielded high match rates to ta

222、rgeted therapies and clinical trials,and also underscored the value of integrating and contextualizing clinical and molecular data to provide physicians with distilled information regarding their patients disease and potentially actionable characteristics.In sum,our Platform demonstrated an ability

223、to help maximize personalized therapeutic options for a broader proportion of patients with cancer,which typically cannot be attained through smaller tumor-only DNA-seq panels.In addition,in a paper we published in Nature Precision Oncology in July 2021,we highlighted the benefits of performing both

224、 solid tumor and liquid biopsy profiling.We observed that the concordance of the results of tissue sequencing and liquid testing,even when concurrently profiled,was approximately 70%at most,with both liquid testing and tissue sequencing missing a selected number of potentially actionable mutations.Y

225、et when both are performed,as Tempus often does,the coverage of potentially actionable mutations increases.We believe the market is recognizing the value of our products and their benefits,as they relate to sequencing both somatic and germline variants,running both solid tumor and liquid biopsies,br

226、oadly sequencing RNA in addition to DNA,making available raw files and structured clinical data,and matching the results to clinical data for the patient sequenced.As a result,our clinical volume in oncology rose from approximately 31,000 samples sequenced in 2018 to approximately 350,000 samples in

227、 2024.Our Neuropsychiatry TestsWe entered neuropsychiatry in 2019.We currently offer our proprietary nP assay for pharmacogenomic testing for patients with psychiatric conditions,such as depression,general anxiety disorder,bipolar disorder,and other relevant diagnoses.Despite the growing prevalence

228、of depression and anxiety,their treatment remains largely the same as it has been for decades.Today,there are dozens of antidepressants that are often prescribed based on trial and error,where psychiatrists alter the dose and class of medications when one fails to work.The difficulties in prescribin

229、g medications leads many patients to take the wrong medications,in the wrong dose.Emerging evidence demonstrates that there are molecular mechanisms that suggest one drug,or class of drugs,may work better than another based on the genetic profile of the patient,and our assay is designed to elucidate

230、 these differences.The following table describes our nP assay.25 Tempus|nP2019 Pharmacogenomic profiling for patients with psychiatric conditions;primarily used for depression Covers 13 validated genes with known roles in pharmacokinetics,pharmacodynamics,and immune response to FDA approved medicati

231、ons that may be prescribed in the neuropsychiatric space Uses matrix-assisted laser desorption ionization-time of flight(MALDI-TOF)mass spectrometry to analyze 80 single nucleotide and small insertion-deletion(indel)variants in the 13 genes.Concurrently,DNA fragment analysis is used to analyze copy

232、number variants in CYP2D6 and a large indel in the SLC6A4 promoter In an effort to bring AI to this field,we are not only performing pharmacogenomic profiling,we also regularly collect two additional data modalities:(i)time on therapy data from the patients EHR(or directly from the ordering physicia

233、n),and(ii)patient reported outcome,or PRO,data through our TempusPRO mobile application.TempusPRO is our patient-facing mobile application that collects PRO measurements on a longitudinal basis.We are also capturing passive lifestyle measurements through mobile sensory devices,such as daily steps an

234、d minutes spent exercising.These measurements serve as a quantitative,unbiased backbone to the more qualitative and subjective measures that are commonplace in psychiatry.As we continue to advance the field of psychiatric medicine,we believe our Platform is well suited to extend to additional neurol

235、ogical conditions beyond depression,anxiety,and bipolar disorder.Data and ServicesOur Data and Services product line facilitates drug discovery and development for life sciences companies through two primary products:Insights and Trials.We also maintain a growing tumor-derived biological modeling(or

236、 organoid)laboratory,which allows us to provide modeling and screening services to our pharmaceutical and biotech clients.One way we measure our data business is based on the remaining total contract value,or the Remaining TCV,that is contractually committed to be delivered in the future.As of Decem

237、ber 31,2024,we have signed contracts with a Remaining TCV of more than$940.0 million,which includes approximately$300.0 million in additional potential future contractual opt-ins.Remaining TCV is equal to the total potential value of signed contracts and assumes the exercise of all contract options,

238、all discretionary opt-ins,and no early termination.Remaining TCV includes the total potential value of the companys strategic collaborations with AstraZeneca AB,or AstraZeneca,and GlaxoSmithKline,or GSK,which,although listed under the Data and Services product line,could be satisfied by the purchase

239、 of any of the companys products and services.Remaining TCV excludes any revenue recognized to date on these contracts or any future adjustments made to the contractual value as a result of amendments or terminations.Our agreements contain termination clauses,including the ability of our counterpart

240、y to terminate for convenience,and there can be no guarantee that contracts will not be terminated,that contractual options and discretionary opt-ins will be exercised,or that we will achieve the full amount of potential revenue represented by these contracts.Remaining TCV is not a calculation of re

241、venue and should be viewed independently of revenue and deferred revenue,as Remaining TCV is not intended to be combined with or replace these items.Similarly,Remaining TCV is not a forecast of future revenue,which can be impacted by,among other things,contract start and end dates,our ability to mee

242、t performance obligations,and the exercise of contractual options or termination rights.Moreover,Remaining TCV may differ from similarly titled metrics presented by other companies and may not be comparable to such other metrics.InsightsHistorically,the primary means for pharmaceutical and biotechno

243、logy companies to build a dataset for drug discovery and development was to run a preclinical study or clinical trial,and to leverage limited datasets such as medical claims data.We believe Tempus is changing the existing paradigm.We launched our Insights product to allow researchers to access large

244、 amounts of multimodal healthcare data that historically did not exist at scale in a single consolidated database.We have amassed a large connected dataset,which we organize in near-real time across multiple modalities and multiple disease areas,allowing us to work with pharmaceutical and biotechnol

245、ogy companies across the drug lifecyclefrom discovery,research and development,and,ultimately,commercialization.26For our Insights offering,we license libraries of linked,de-identified clinical,molecular,and imaging data,and provide a suite of analytic and cloud-and-compute tools for discovery,resea

246、rch,development,and other commercial purposes.Our primary customers are pharmaceutical and biotechnology companies.These customers either pay us on a per file basis or through multi-year data licensing agreements to use our de-identified patient database.We currently work with 19 of the 20 largest p

247、ublic pharmaceutical companies based on 2023 revenue.We believe we offer a unique value proposition to the industry as a cost-effective source of high-quality and comprehensive data on targeted patient populations.Our data is useful across the oncology drug development value chain,and our biotechnol

248、ogy and pharmaceutical customers are using the data to inform decisions in a variety of discovery and development applications,selected below.One metric that illustrates the utility of our data to our customers is“Net Revenue Retention.”Net Revenue Retention compares the annual Insights product reve

249、nue generated from all customers that made an Insights purchase in one year to the annual Insights product revenue generated from the same cohort of customers in the subsequent year.Net Revenue Retention is not a calculation of revenue and should be viewed independently of revenue and deferred reven

250、ue,as Net Revenue Retention is not intended to be combined with or replace these items.Similarly,Net Revenue Retention is not a forecast of future revenue.Moreover,Net Revenue Retention may differ from similarly titled metrics presented by other companies and may not be comparable to such other metr

251、ics.For the year ended December 31,2024,Net Revenue Retention was approximately 140%compared to the same cohort of customers for the period ended December 31,2023.To illustrate an example of how our data can be applied,in December 2020,we published a peer-reviewed study in ScienceDirect in which we

252、analyzed longitudinal real-world data,or RWD,from a large cohort of patients with breast cancer(n=4,000)to test whether results were consistent with previous clinical studies and to demonstrate the real-world evidence validity of our database.We also evaluated whole-transcriptome sequencing as a com

253、plementary diagnostic tool(n=400).The conclusions of the study demonstrated that our database mirrored the overall population of patients with breast cancer in the United States,and that near real-time,RWD analyses are feasible in a large,highly heterogeneous database.Furthermore,the study demonstra

254、ted that molecular data may aid deficiencies and discrepancies observed from breast cancer clinical RWD.Because many of our data profiles regularly update with clinical outcome and response data over time,the utility of a single de-identified record may increase over time.As such,the files we genera

255、te by sequencing a patient,when connected to clinical data,are valuable to pharmaceutical and biotechnology companies,as they not only allow users to gain molecular insight into what is happening among cohorts of patients,but they also allow users to track those cohorts over time.As a result,our fil

256、es behave as if they have a“lifetime value”that has the potential to increase over time,in a manner similar to a content company where you pay to create content and then monetize the content over time as people subscribe to access the content.27 TrialsTrials is our second offering within our Data an

257、d Services product line and leverages our broad network of physicians we work with in oncology to provide clinical trial matching services for pharmaceutical companies trying to reach hard-to-find and underserved patient populations.Our clinical trial matching product is built on top of our near rea

258、l-time data feeds and harnesses AI to accelerate the connection between patients,clinical trial sites(hospitals)and clinical trial sponsors(life sciences companies).We empower both oncologists to help their patients find clinical trials and pharmaceutical companies to enroll patients into their tria

259、ls.We generate revenue from both matching the patient to the trial(through notices we send to physicians alerting them of potential trials that are a fit for their patients),and from the patient actually enrolling in the trial.Our Trials product is a bold initiative that we do not believe has been i

260、mplemented at scale in the United States by any other organization.We are endeavoring to create a just-in-time network across a wide variety of academic medical centers and community providers,that can support hundreds or even thousands of trials,in which the administrative and logistical foundation

261、 is uniform across the entire network.This network allows us to identify a patient that is a match for a targeted trial and get that patient enrolled within days,even if the trial was not previously open at the hospital(assuming consent of the trial sponsor),anywhere in the United States.Prior to Te

262、mpus,we believe it would have been virtually impossible to even attempt to build this type of just-in-time program across oncology,as the required ingredients for success are unique to our Platform,namely:(i)a large genomic sequencing business that is widely adopted and allows for the identification

263、 of patients that are molecular matches to trials;(ii)the ability to structure clinical data for those patients in near real time to filter for inclusion and exclusion criteria;(iii)direct pipelines allowing data to be transferred to and from the laboratory and provider;and(iv)an analytic engine abl

264、e to stratify patients and follow each unique patient journey ensuring that patients actually enroll in the studies.Our clinical trial matching offering is called the TIME Trial program,which we launched in June of 2019.Since its introduction,this program has gained significant traction with more th

265、an 250 clinical trials signed into the network.More than 30,000 patients were identified for potential enrollment into clinical trials in our network as of December 31,2024.We believe the breadth of our network,the data to which we have near real-time access,and our relationships with oncologists en

266、able us to offer a clinical trial matching service that has the potential to materially expand patient access to and accelerate enrollment in clinical trials in the United States.28One of the primary benefits of our Trials product is our ability to facilitate the initiation of a clinical trial in a

267、new location in a short amount of time.Third-party research suggests that it takes 6-12 months,on average,to initiate a new trial site for an ongoing clinical trial in the United States.We have been able to substantially streamline this process by leveraging technology and introducing a standard met

268、hodology,with activation of new sites through our Trials product taking approximately two weeks on average in 2024.A comparison of our average time from site initiation to patient consent with the industry average is below:In addition to TIME,we provide other clinical trial services and conduct our

269、own studies as part of our Trials program,all with a goal of identifying new therapies and bringing them to market more efficiently.In January 2022,we acquired Highline Consulting,LLC,a contract research organization,or CRO,which we subsequently renamed Tempus Compass,LLC,or Tempus Compass.Tempus Co

270、mpass manages and executes early and late-stage clinical trials,primarily in oncology.We also partner with life sciences companies to sponsor studies of drugs,devices,and diagnostics,integrating our life science solutions to help bring new drugs to market faster.Each of the products and services wit

271、hin our Trials program complement each other to create a suite of integrated solutions for life sciences companies from early discovery to commercialization.Tumor Derived Biological ModelingOrganoidsIn addition to our efforts to collect vast amounts of phenotypic,morphologic,and molecular data,we ha

272、ve built a large,biological modeling lab that allows us to test various theories in vitro through our large repository of tumor-derived Organoids,and to perform drug screening for our various life sciences clients.Many of our Organoids are fully characterized and sequenced using our NGS panels,provi

273、ding genomic and transcriptomic data for our models,allowing us to explore various hypotheses that enhance our data.Examples of hypotheses we are able to test in our Organoid lab include:(i)which therapeutics are most effective;(ii)differential levels of drug response by tumor type,genomic profile,o

274、r other targeted attributes;(iii)discovery of RNA signatures;(iv)attributes of responders and non-responders;and(v)response rates in therapy-resistant models.We work with numerous collaborators including biotechnology companies,pharmaceutical companies,academic institutions,and government labs.Since

275、 2017,we have scaled our sample collection efforts and have received approximately 4,100 tumor samples to date.These samples cover a wide range of cancer subtypes,allowing us to work on comprehensive drug screening applications across multiple epithelial based tumor types,such as breast,lung,colorec

276、tal,and pancreatic.One of the goals of this screening is to predict a series of therapeutic responses in our Organoids and then test whether or not patients are experiencing similar responses in the clinical setting.29We view biological models as another form of data.Our efforts to grow Organoids ar

277、e part of our overall strategy to leverage the best of systems biology along with the best of AI to collect the requisite data needed to produce answers broadly throughout healthcare.AI ApplicationsThe vastness of our dataset,along with our connected platform,creates an opportunity to use data to al

278、gorithmically diagnose and treat patients.Our third product line,AI Applications,or Algos,is focused on developing and providing diagnostics that are algorithmic in nature,implementing new software as a medical device,and building and deploying clinical decision support tools.The primary product of

279、AI Applications is currently“Next,”an AI platform that leverages machine learning to apply an“intelligent layer”onto routinely generated data to proactively identify and minimize care gaps for oncology and cardiology patients.As this product gains adoption,we intend to leverage large language models

280、,generative AI algorithms,and our vast database of de-identified data to develop algorithmic diagnostics designed to identify these patients earlier in their disease progression,when treatments are most effective.For example,algorithmic diagnostics that integrate multimodal data can be used to creat

281、e a more accurate risk profile for patients,leading to improved outcomes and reduced cost.Our repository of multimodal data allows us to find associations and patterns that are largely invisible through a single data modality,but readily apparent when combined.In addition,we find the strength of our

282、 analytic models,and our ability to deploy them clinically,improves as we add additional datasets.While we plan to continue developing our own proprietary software and algorithms,from time to time,we also utilize open source technologies or in-license technologies from third parties.The example diag

283、ram below represents how algorithm-based diagnostics work and the value of multimodal data as it relates to improved analytics:30Algorithm-based diagnostics are already being used in healthcare,but are not widespread.For example,algorithms exist today that leverage EHR data and lab results to predic

284、t early onset of hospital-borne infections,but these tools are still in the very early stages of adoption and validation.While Algos today represent only a small proportion of the diagnostics market,we expect their adoption to grow substantially in the future.We believe Algos represent a significant

285、 long-term opportunity that may be substantially larger than our other existing product lines.We believe our ability to launch Algos at scale is a key differentiator of our Platform.Our Oncology Algos PortfolioWe believe our robust,multimodal dataset creates an opportunity for Algos that otherwise w

286、ould not be possible and allows us to build AI models at scale,clinically validate them,and deploy the resulting Algos into clinical practice.We currently offer a suite of Algos in oncology,and have more in various stages of development.As of December 31,2024,more than 80,000 molecular oncology Algo

287、s have been ordered with our various genomic assays.Most of the Algos we currently offer are part of our xR assay,and we do not bill separately for them.Some Algos will likely yield little to no reimbursement until their clinical utility is established or will be ordered separately with our existing

288、 NGS assays or diagnostics to enhance the actionable information for physicians,and some may obtain reimbursement at prevailing rates for comparable tests.Algo Launch Year Description Oncology Tumor Origin(“TO”)Test2021 Predicts the site of origin for cancer patients whose primary tumor site is unkn

289、own using tumor RNA expression results Intended use of the TO test is for cancers of unknown primary,or CUPs,and may help clinicians make more informed decisions where other clinical information like imaging and immunohistochemistry results do not provide a definitive diagnosis Uses information from

290、 analysis of nucleic acids by NGS performed as part of a separately ordered genomic or transcriptomic test Built using a large internal database of more than 20,000 annotated tumors with transcriptomic molecular data.By comparing the molecular profile(transcriptome)of the patients cancer with profil

291、es of other cancers in our database,we can help pinpoint the origin of the patients cancer,potentially helping to inform the course of therapy For the year ended December 31,2024,ordered on approximately 10%of our solid tumor profiles Homologous Recombination Deficiency(“HRD”)Test 2020 A DNA-based a

292、lgorithmic test that helps identify if a patient has HRD,providing a comprehensive view into a patients ability to repair double-stranded DNA breaks HRD status can be used to identify patients who may be sensitive to PARP inhibitors and/or platinum-based chemotherapy Takes into account results from

293、our solid tumor profiling,giving a full view into commonly mutated genes in the HR-pathway,along with a genome wide l LOH score,giving a clinician a complete view of HRD status Can be ordered across all major cancer subtypes and does not require additional tissue from the patient Currently incorpora

294、ting RNA into a second version of the algorithm,which is intended to improve prediction 31Algo Launch Year Description Dihydropyrimidine Dehydrogenase Deficiency(“DPYD”)Test 2021 Identifies certain alterations in the DPYDgene,which may be associated with a patients potential toxicity to 5-FU/Capecit

295、abine chemotherapy based on the associated drug labeling and guidelines from the Clinical Pharmacogenomics Implementation Consortium,or CPIC.Provides insight into the potential likelihood of a patient developing severe or even fatal toxicity of 5-FU/Capecitabine chemotherapy by covering five SNVs in

296、 DPYD genes,providing a more complete patient profile.According to CPIC,5-7%of patients test positive for DPYD deficiency and should be considered for monitoring or dose reduction.This algorithm uses sequencing data generated as a part of a separately-ordered Tempus|xT Solid Tumor+Normal test.Tempus

297、 DPYD is available pan-cancer although it is most relevant in colorectal,breast,pancreatic and GI cancer patients who are being considered for treatment with 5-FU/Capecitabine chemotherapy.Tempus Purist2023 Tempus Purist test is an algorithm that classifies pancreatic ductal adenocarcinomas(PDAC)pat

298、ients into one of two subtypes(basal-like or classical).Patients with the basal subtype have a worse prognosis and are less likely to benefit from FOLFIRINOX therapy than classical patients.Uses information from nucleic acids by NGS performed as part of a separately ordered genomic or transcriptomic

299、 test.USES a k-top scoring pair(k-TSP)method(8 top scoring pairs,16 genes in total)to assign a basal probability score.Patients with a basal probability score of 50 are categorized as basal subtype,while those withbasal probability score 50 are categorized as classical subtype.Our Cardiology AlgosHe

300、art disease is the leading cause of death in the United States.About 700,000 Americans die from heart disease annually,with more than 11%of American adults diagnosed with heart disease and millions of patients suffering from undiagnosed,life-threatening,yet highly treatable conditions such as AFib,c

301、ardiomyopathy,and valvular heart disease,to name a few.Tempus is working on solutions to find,diagnose,and help treat these patients earlier in order to improve patient outcomes,using routinely generated clinical data,such as data from a 12-lead ECG,a widely used and easily acquired medical test tha

302、t measures the electrical activity of the heart,to screen patients who might be at high risk and help navigate them to the appropriate interventional therapy.In cardiology,we ingest multimodal data and use approximately 60 algorithms to identify potential care gaps and continuously monitor patient d

303、ata to find at-risk patients who may be falling through a care gap unbeknownst to their physician,and automatically notify care teams of any needed follow-up or disease progression.More than 100 hospitals nationwide are currently powered by Tempus Next and more than 45,000 patients are screened per

304、month.SMSM 32We are also developing algorithmic models that aid clinicians in identifying patients at increased risk of developing atrial fibrillation,or AFib,along with a variety of other cardiac conditions.These Algos are trained using a de-identified subset from approximately 3.5 million ECGs,acr

305、oss more than 800,000 patients,with decades of longitudinal clinical data,including outcome and response data.The FDA granted Tempus breakthrough status for our first ECG software device,which employs a diagnostic algorithm designed to identify patients at high risk of developing AFib in certain pop

306、ulations(patients 40 years of age and older,without pre-existing or concurrent AFib or atrial flutter,and who are at elevated risk of stroke based on a commonly used clinical stroke risk assessment tool(i.e.,CHA2DS2-VASc score of 4).Algo Launch Year Description Cardiology Atrial Fibrillation Test202

307、3(in clinical trial setting)We have developed an algorithm designed to predict AFib from a normal ECG for certain populations.About 3.5%of patients who receive ECGs appear not to have AFib but will develop AFib,acute coronary syndrome,or similar condition within one year.This Algo is designed to pre

308、dict major cardiac trauma and stroke risk from these normal ECG results.The Tempus AFib test received FDA breakthrough designation in March 2021 for patients 40 years of age and older,without pre-existing or concurrent AFib or atrial flutter,and who are at elevated risk of stroke based on a commonly

309、 used clinical stroke risk assessment tool(i.e.,CHA DS-VASc score of 4).We are also advancing Algos that are designed to predict aortic stenosis,and we are working on other disease areas within cardiology,such as low ejection fraction and familial hypercholesterolemia.We are also advancing Algos tha

310、t are designed to predict aortic stenosis,and we are working on other disease areas within cardiology,such as low ejection fraction and familial hypercholesterolemia.If broadly deployed,we believe these Algos could have widespread clinical applicability,increase life expectancy,and reduce the total

311、cost of care.In addition to algorithms based on NGS testing or in the cardiology space,we currently offer more than 50 algorithms and are continuing to develop additional algorithms derived from radiologic images and digital pathology slides.In October 2022,we acquired Arterys,Inc.,a company that pr

312、ovides a platform to derive insights from radiologic medical images to improve diagnostic decision-making,efficiency,and productivity across multiple disease areas.We have also developed algorithms based on Immunohistochemistry,or IHC,and H&E staining,which can be used,among other things,to help ide

313、ntify patients who may be eligible for additional treatments or clinical trials.CommercializationOur commercial efforts are generally focused on driving increased adoption of our various products and services,both by increasing the utilization of existing customers and securing new customers.We empl

314、oy targeted sales and business development organizations,whose team members are engaged in direct sales and marketing efforts.Our commercial teams typically target healthcare providers and life sciences companies,which are the main purchasers of our products and services.We describe below our overal

315、l commercial strategy for each of our three products.GenomicsOur Genomics product line,largely made up of molecular testing,has two primary customers:physicians and bio-pharma companies.When we sell our tests to physicians we are typically providing them as part of routine clinical care and we are o

316、ften billing insurance and seeking reimbursement on behalf of the patients for whom the test was ordered.When we sell our test to bio-pharma,we are typically being paid as a contract sequencing provider,either for the trials they are running or as a companion diagnostic to their drug.On the physicia

317、n side,we commercialize our Genomics products in the United States to clinicians and healthcare providers largely through our dedicated clinical sales organization,that calls on individual 22 33doctors or medical practices.As of December 31,2024,our clinical sales organization in the United States i

318、ncluded approximately 210 sales representatives who are primarily contacting oncologists,psychiatrists,and other healthcare providers.Our sales representatives typically have backgrounds either in a particular disease area(such as oncology or neuropsychiatry)or in laboratory testing and therapeutics

319、 more generally.We supplement our commercial team with clinical specialists with extensive medical affairs experience who provide molecular support in the field.In oncology,which currently is our largest market,we are focused on driving adoption by targeting individual treating physicians,academic m

320、edical centers,community oncology practices,leading physician networks,and industry associations.We also are exploring relationships with third-party payers and governmental institutions.We have a land and expand strategy,by account,whereby we attempt to sign new accounts and increase adoption of ou

321、r platform within these accounts over time.As such,we often begin a relationship that is transactional in nature,but seek over time,to work on a more comprehensive basis with healthcare providers,serving an ever increasing percentage of our molecular diagnostic needs over time.We find that once a ph

322、ysician starts using Tempus,if they order more than 5 oncology NGS tests from us,their 12-month retention rate is 87%.In addition,we believe that interactions among treating physicians help drive adoption of our products.We are focused on key opinion leaders in the industry through direct outreach a

323、nd indirect marketing efforts.As of December 31,2024,we have either published or been acknowledged in 600 publications(more than 450 Tempus-authored),including the following:163 total(120 Tempus-authored)peer-reviewed articles published or accepted for publication in major journals,including publica

324、tions such as Nature Biotechnology,Clinical Breast Cancer,Nature Medicine,and Cell.339 total(281 Tempus-authored)poster presentations based on clinical and research data that have been accepted and presented at major scientific conferences.37 oral presentations at scientific meetings such as the ASC

325、O,ASCO Gastrointestinal and Genitourinary Cancer Symposiums,San Antonio Breast Cancer Symposium,and the American Heart Association Scientific Sessions.We have a similar strategy in neuropsychiatry,in which we aim to increase the commercial adoption of our nP test for depression as part of the rapidl

326、y growing market for pharmacogenomic testing,with a goal to better understand,diagnose and treat neuropsychiatric disorders.Our commercial strategy for other disease areas is expected to follow our strategy in oncology,which is to focus on offering a broad range of molecular diagnostics to the marke

327、t,that are connected to clinical data,so we can track how molecular results correlate with outcomes and responses,thereby making our tests smarter and more personalized overtime.Research TestingA small component of our genomic testing involves testing performed in a research capacity.This type of te

328、sting is typically done under an agreed upon contracted arrangement for specific tests at specific prices and volumes.Typical customers in these arrangements are pharmaceutical companies engaged in testing for clinical trials,researchers who need genomic testing to further research activities,or a c

329、ompany marketing products or services of their own who elects to use us as a reference laboratory.In this type of research testing,the agreed upon rate for testing may vary significantly,and in some cases may even be offered as an in-kind service in exchange for other rights we obtain in the contrac

330、ted relationship.As it relates to selling our Genomic Products to bio-pharma,we have a dedicated team of sales executives focused on calling on biotech and pharmaceutical companies who use genomic sequencing services predominantly for the research they are conducting,the clinical trials they are run

331、ning,or as a companion diagnostic to the extent their therapeutic relies on a bio-marker.To this group,we are typically selling retrospective and prospective sample testing services,as well as companion diagnostic development to support the approval and commercialization of therapeutics.Data and Ser

332、vicesIn addition to our field sales force,our Data and Services products rely on a dedicated business development team focused on enterprise sales to pharmaceutical and biotechnology companies in the United States and abroad.Our strategy with each customer is to demonstrate the value proposition of

333、our Platform and de-identified datasets,and to expand the utilization of our Data and Services products across the organization from early-stage research through clinical development 34to commercialization.Given the broad and differentiated utility of our Platform,we believe we can support our pharmaceutical and biopharmaceutical customers across many applications,including:early stage research an