Caris Life Sciences Publishes Study Showing its Multi-Layer AI-Based Tissue of Origin Predictions are Best-in-Class and Identify when Patients have been Misdiagnosed
Caris Life Sciences (NASDAQ: CAI) has published a groundbreaking study on its enhanced Caris GPSai鈩� platform in AACR's Cancer Research Communications Journal. The deep learning AI system, trained on over 200,000 cases, demonstrates exceptional accuracy in identifying tumor tissue origin, achieving 95.0% accuracy in non-CUP cases.
The system successfully identified tissue origin in 84.0% of CUP and 96.3% of non-CUP cases during validation across 97,820 cases. In clinical use, GPSai led to diagnosis changes in 704 patients, with 86.1% of cases impacting treatment eligibility based on Level 1 clinical evidence. Notably, 53.6% of surveyed physicians modified treatment plans based on these findings.
Caris Life Sciences (NASDAQ: CAI) ha pubblicato uno studio innovativo sulla sua piattaforma avanzata Caris GPSai鈩� nella rivista Cancer Research Communications dell'AACR. Il sistema di intelligenza artificiale basato su deep learning, addestrato su oltre 200.000 casi, dimostra un'accuratezza eccezionale nell'identificazione dell'origine del tessuto tumorale, raggiungendo un'accuratezza del 95,0% nei casi non CUP.
Durante la validazione su 97.820 casi, il sistema ha identificato con successo l'origine del tessuto nel 84,0% dei casi CUP e nel 96,3% dei casi non CUP. Nell'uso clinico, GPSai ha portato a cambiamenti diagnostici in 704 pazienti, con un impatto sul trattamento basato su evidenze cliniche di Livello 1 nel 86,1% dei casi. In particolare, il 53,6% dei medici intervistati ha modificato i piani terapeutici in base a questi risultati.
Caris Life Sciences (NASDAQ: CAI) ha publicado un estudio innovador sobre su plataforma mejorada Caris GPSai鈩� en la revista Cancer Research Communications de AACR. El sistema de inteligencia artificial de aprendizaje profundo, entrenado con m谩s de 200,000 casos, demuestra una precisi贸n excepcional para identificar el origen del tejido tumoral, alcanzando una precisi贸n del 95,0% en casos no CUP.
El sistema identific贸 con 茅xito el origen del tejido en el 84,0% de los casos CUP y en el 96,3% de los casos no CUP durante la validaci贸n en 97,820 casos. En uso cl铆nico, GPSai llev贸 a cambios en el diagn贸stico en 704 pacientes, con un impacto en la elegibilidad para tratamiento basado en evidencia cl铆nica de Nivel 1 en el 86,1% de los casos. Notablemente, el 53,6% de los m茅dicos encuestados modificaron sus planes de tratamiento bas谩ndose en estos hallazgos.
Caris Life Sciences (NASDAQ: CAI)電� AACR鞚� Cancer Research Communications 鞝�雱愳棎 鞛愳偓鞚� 頄レ儊霅� Caris GPSai鈩� 頂岆灚韽检棎 甏�頃� 須嶊赴鞝侅澑 鞐瓣惮毳� 氚滍憸頄堨姷雼堧嫟. 20毵� 瓯� 鞚挫儊鞚� 靷搿� 頃欖姷霅� 霐ル煬雼� AI 鞁滌姢韰滌潃 膦呾枒 臁办 旮办洂鞚� 鞁濍硠鞐愳劀 霙办柎雮� 鞝曧檿霃勲ゼ 氤挫棳欤茧┌, 牍�-CUP 靷鞐愳劀 95.0%鞚� 鞝曧檿霃�毳� 雼劚頄堨姷雼堧嫟.
鞚� 鞁滌姢韰滌潃 97,820瓯挫潣 靷 瓴�歃濎棎靹� CUP 靷鞚� 84.0%鞕赌 牍�-CUP 靷鞚� 96.3%鞐愳劀 臁办 旮办洂鞚� 靹标车鞝侅溂搿� 鞁濍硠頄堨姷雼堧嫟. 鞛勳儊 靷毄 鞁� GPSai電� 704氇呾潣 頇橃瀽 歆勲嫧 氤�瓴届潉 鞚措亴鞐堨溂氅�, 鞚� 欷� 86.1%電� 1雼硠 鞛勳儊 攴缄卑鞐� 霐半ジ 旃橂 鞝來暕靹膘棎 鞓來枼鞚� 氙胳长鞀惦媹雼�. 韸鬼瀳, 53.6%鞚� 靹る鞐� 彀胳棳頃� 鞚橃偓霌れ澊 鞚� 瓴瓣臣毳� 氚旐儠鞙茧 旃橂 瓿勴殟鞚� 靾橃爼頄堨姷雼堧嫟.
Caris Life Sciences (NASDAQ : CAI) a publi茅 une 茅tude r茅volutionnaire sur sa plateforme am茅lior茅e Caris GPSai鈩� dans la revue Cancer Research Communications de l'AACR. Le syst猫me d'IA par apprentissage profond, entra卯n茅 sur plus de 200 000 cas, d茅montre une pr茅cision exceptionnelle dans l'identification de l'origine des tissus tumoraux, atteignant une pr茅cision de 95,0 % pour les cas non CUP.
Le syst猫me a r茅ussi 脿 identifier l'origine des tissus dans 84,0 % des cas CUP et 96,3 % des cas non CUP lors de la validation sur 97 820 cas. En utilisation clinique, GPSai a conduit 脿 des changements de diagnostic chez 704 patients, avec un impact sur l'茅ligibilit茅 au traitement bas茅 sur des preuves cliniques de niveau 1 dans 86,1 % des cas. Notamment, 53,6 % des m茅decins sond茅s ont modifi茅 leurs plans de traitement en se basant sur ces r茅sultats.
Caris Life Sciences (NASDAQ: CAI) hat eine bahnbrechende Studie zu seiner verbesserten Caris GPSai鈩�-Plattform im AACR-Fachjournal Cancer Research Communications ver枚ffentlicht. Das Deep-Learning-KI-System, trainiert an 眉ber 200.000 F盲llen, zeigt eine au脽ergew枚hnliche Genauigkeit bei der Identifizierung des Tumorgewebsursprungs und erreicht eine Genauigkeit von 95,0% bei Nicht-CUP-F盲llen.
Das System identifizierte w盲hrend der Validierung an 97.820 F盲llen erfolgreich den Gewebeursprung in 84,0% der CUP- und 96,3% der Nicht-CUP-F盲lle. In der klinischen Anwendung f眉hrte GPSai bei 704 Patienten zu Diagnosen盲nderungen, wobei in 86,1% der F盲lle die Behandlungsoptionen basierend auf Level-1-Kliniknachweisen beeinflusst wurden. Bemerkenswert ist, dass 53,6% der befragten 脛rzte ihre Behandlungspl盲ne aufgrund dieser Ergebnisse angepasst haben.
- Achieved 95.0% accuracy in identifying tumor tissue origin in non-CUP cases
- Successfully identified tissue origin in 84.0% of CUP and 96.3% of non-CUP cases
- Changed diagnosis in 704 patients with 86.1% impact on treatment eligibility
- 53.6% of surveyed physicians changed treatment plans based on findings
- Platform trained on extensive dataset of over 200,000 cases
- None.
Insights
Caris GPSai's deep learning tool shows 95% accuracy identifying cancer origins, directly changing patient diagnoses and treatment paths.
Caris Life Sciences has published validation data for their enhanced Caris GPSai鈩� diagnostic tool in AACR's Cancer Research Communications Journal, demonstrating significant clinical impact. This deep learning AI system, trained on comprehensive molecular data from over 200,000 cases, classifies tumors into 90 categories with remarkable precision 鈥� 95.0% accuracy in non-CUP (Cancer of Unknown Primary) cases.
The most compelling aspect is the real-world clinical validation. During an eight-month implementation period, GPSai actually changed the diagnosis in 704 patients, with these changes supported by orthogonal evidence like imaging and molecular markers. These weren't merely academic reclassifications 鈥� they impacted treatment eligibility in 86.1% of cases based on Level 1 clinical evidence, with 53.6% of physicians surveyed changing treatment plans accordingly.
The technology represents a significant advance from traditional machine learning to deep learning approaches. By leveraging whole exome and whole transcriptome sequencing (WES/WTS), it can identify potential misdiagnoses during routine molecular profiling without requiring additional tissue samples 鈥� a critical advantage in oncology where tissue availability is often limited.
The clinical case example cited is particularly striking: a patient initially diagnosed with triple-negative breast cancer was correctly identified as having B-cell lymphoma, which would completely change treatment approach. For CUP patients specifically, who historically face poor outcomes due to diagnostic uncertainty, the tool successfully reported a tissue of origin in 84.0% of cases, potentially bringing these patients into standard treatment protocols with improved outcomes.
Caris' AI diagnostic tool demonstrates clinical utility and market differentiation by changing diagnoses and treatment paths for misdiagnosed cancer patients.
Caris Life Sciences has published compelling validation data for its AI-powered diagnostic tool that creates real differentiation in the precision oncology market. The enhanced Caris GPSai鈩� demonstrates 95% accuracy in tumor origin identification, addressing two critical unmet needs: identifying primary sites for Cancers of Unknown Primary (CUP) and catching misdiagnosed cancers during routine molecular profiling.
The commercial potential hinges on three key metrics. First, the ability to report a tissue of origin in 84.0% of CUP cases represents a significant advance for these traditionally difficult-to-treat patients. Second, during eight months of clinical implementation, the tool changed diagnoses in 704 patients, with 86.1% of these changes impacting treatment eligibility based on Level 1 evidence. Third, and perhaps most compelling for adoption, 53.6% of surveyed physicians reported changing treatment plans based on these findings.
This technology strengthens Caris' competitive position in the molecular diagnostics market. By leveraging its extensive database of over 200,000 profiled cases for AI training, the company is creating high barriers to entry. The integration of whole exome and whole transcriptome sequencing with deep learning capabilities represents a shift from earlier machine learning approaches, potentially making competing offerings obsolete.
For payers, this tool may improve the cost-effectiveness of precision oncology by ensuring patients receive appropriate targeted therapies based on accurate diagnoses, rather than treatments for incorrectly identified cancers. The case example of reclassifying a presumed triple-negative breast cancer as B-cell lymphoma illustrates how this technology can substantially alter treatment pathways and potentially improve outcomes while avoiding ineffective therapies.
Caris GPSai鈩� utilizes deep learning to significantly improve diagnostic accuracy for cancers of unknown primary and misclassified tumors
This latest advancement marks a shift from traditional machine learning to a deep learning-based approach, enabling more precise prediction of tumor tissue of origin and the identification of potential misdiagnoses during routine molecular profiling, ultimately supporting more informed treatment decisions and potentially improving patient outcomes.
The enhanced Caris GPSai was trained on WES/WTS data from over 200,000 Caris-profiled cases and classifies tumors into 90 categories.听The tool demonstrated
"The latest version of GPSai, which is a part of Caris' comprehensive molecular profiling platform, represents a major advancement in precision oncology," said , Caris SVP and Chief Clinical Officer and Pathologist-in-Chief. "By leveraging deep learning and whole exome and whole transcriptome sequencing, GPSai enhances diagnostic confidence, enabling more accurate identification of primary tumor sites and supporting more personalized treatment decisions, without the need for additional tissue samples."
In clinical use over eight months, GPSai changed the diagnosis in 704 patients, supported by orthogonal evidence such as imaging and molecular markers. These diagnostic shifts impacted treatment eligibility in
"By enhancing diagnostic accuracy, GPSai empowers physicians to make more informed treatment decisions and identify the tumor type for patients with CUP and those that have been misdiagnosed," said , President of Caris. "For example, we had a case where we found a woman diagnosed with triple negative breast cancer, who actually had B-cell Lymphoma; getting the correct diagnosis had a profound effect on her life."
The publication can be viewed in its entirety on the .
About Caris Life Sciences听
Caris Life Sciences庐鈥�(Caris) is a leading, patient-centric, next-generation AI TechBio company and precision medicine pioneer that is actively developing and commercializing innovative solutions to transform healthcare. Through comprehensive molecular profiling (Whole Exome and Whole Transcriptome Sequencing) and the application of advanced AI and machine learning algorithms at scale, Caris has created the large-scale, multimodal clinico-genomic database and computing capability needed to analyze and further unravel the molecular complexity of disease. This convergence of next-generation sequencing, AI and machine learning technologies, and high-performance computing provides a differentiated platform to develop the latest generation of advanced precision medicine diagnostic solutions for early detection, diagnosis, monitoring, therapy selection and drug development.听
Caris was founded with the belief and vision that combining a vast set of consistently generated molecular information with robust data-driven insights could realize the potential of precision medicine for patients. Headquartered in
Forward Looking Statements
This press release contains forward-looking statements, within the meaning of the federal securities laws, about Caris Life Sciences and its business. All statements other than statements of historical facts contained in this press release are forward-looking statements. In some cases forward-looking statements can be identified by words such as "may," "will," "should," "would," "expect," "plan," "anticipate," "could," "intend," "target," "project," "potential," "contemplate," "believe," "estimate," "predict," "potential," "supports" or "continue" or similar expressions.听听听
You should not rely upon forward-looking statements as predictions of future events.听 Although we believe that the expectations reflected in these forward-looking statements are reasonable based on information currently available to us, we cannot guarantee that the future results, discoveries, levels of activity, performance or events and circumstances reflected in forward-looking statements will be achieved or occur. Forward-looking statements involve known and unknown risks and uncertainties, some of which are beyond our control.听 Risks and uncertainties that could cause our actual results to differ materially from those indicated or implied by the forward-looking statements in this press release include, among other things: developments in the precision oncology industry; future financial performance, results of operations or other operational results or metrics; development, validation and timing of future solutions; the rapidly evolving competitive environment in which we operate; third-party payer reimbursement and coverage decisions; our ability to protect and enhance our intellectual property; regulatory requirements, decisions or approvals (including the timing and conditions thereof) related to our solutions;听 our compliance with laws and regulations; and our ability to hire and retain key personnel as well as risks, uncertainties, and other factors described in the section titled "Risk Factors" and elsewhere in the prospectus for our initial public offering filed with the Securities and Exchange Commission on June 20, 2025, and in our other filings we make with the SEC from time to time.听 We undertake no obligation to update any forward-looking statements to reflect changes in events, circumstances or our beliefs after the date of this press release, except as required by law.听
Caris Life Sciences Media:
Corporate Communications
[email protected]
214.294.5606
View original content to download multimedia:
SOURCE Caris Life Sciences
FAQ
What accuracy rate did Caris GPSai achieve in identifying tumor tissue origin?
How many patients had their diagnosis changed by Caris GPSai?
What percentage of treatment plans were affected by Caris GPSai's findings?
How large was the training dataset for Caris GPSai?
What is the main purpose of Caris GPSai technology?
