MI FOLFOXai™ and MI GPSai™
Caris Life Sciences® continues to be at the forefront of innovation in molecular science and artificial intelligence (AI) with the recent publications of MI FOLFOXai™ and MI GPSai™. These AI-based, molecular signature prediction tools have advanced personalized medicine by providing more diagnostic insight and overall more effective individualized treatments. The results of the studies, as discussed in more detail below, were published in leading industry journals.
Colorectal cancer is the third most common cancer in adults worldwide, with more than 1.9 million new cases diagnosed in 2020 according to the World Health Organization. The choice of treatment is critical to the patient’s prognosis with approximately 25% of colorectal cancer patients presenting with Stage IV – or metastatic – disease, where the cancer has spread to other parts of the body.
Traditionally, FOLFOX, FOLFIRI, or FOLFOXIRI chemotherapy with bevacizumab is considered standard first-line treatment option for patients with metastatic colorectal cancer (mCRC). These standard of care therapies are highly toxic and many patients can only tolerate the therapies for a limited time while physicians look for evidence of therapy benefit.
To help support critical therapeutic decisions for patients with mCRC, Caris Life Sciences® developed and validated a molecular signature predictive of efficacy of oxaliplatin-based chemotherapy combined with bevacizumab (BV). Published in Clinical Cancer Research in December 2020, validation studies showed positive results for MI FOLFOXai™, the company’s AI-based predictor intended to gauge a mCRC patient’s likelihood of benefit from first-line treatment FOLFOX followed by FOLFIRI versus FOLFIRI followed by FOLFOX.
Using two independent data sets, a 67-gene signature was cross-validated in a training cohort (N = 105) which demonstrated the ability of FOLFOXai to distinguish FOLFOX-treated patients with mCRC with increased benefit from those with decreased benefit. The signature was predictive of TTNT and overall survival (OS) in an independent real-world evidence (RWE) dataset of 412 patients who had received FOLFOX/bevacizumab in first line and inversely predictive of survival in RWE data from 55 patients who had received first-line FOLFIRI. Blinded analysis of TRIBE2 samples confirmed that FOLFOXai was predictive of OS in both oxaliplatin-containing arms (FOLFOX HR, 0.629; P = 0.04 and FOLFOXIRI HR, 0.483; P = 0.02). Clinical application of MI FOLFOXai, as a part of routine molecular profiling, could lead to improved outcomes for patients with mCRC and other cancers who are typically candidates for oxaliplatin-based chemotherapy.
The studies demonstrated that the overall OS of patients treated in a manner consistent with the FOLFOXai prediction was 17.5 months longer (71%) than the OS of patients treated counter to the prediction.1 FOLFOXai is the first clinically validated machine-learning powered molecular predictor of chemotherapy efficacy in patients with mCRC.
For more information on MI FOLFOXai™ please visit:
1 Abraham JP, Magee D, Cremolini C, Antoniotti C, Halbert DD, Xiu J, Stafford P, Berry DA, Oberley MJ, Shields AF, Marshall JL, Salem ME, Falcone A, Grothey A, Hall MJ, Venook AP, Lenz HJ, Helmstetter A, Korn WM, Spetzler DB. Clinical validation of a machine-learning derived signature predictive of outcomes from first-line oxaliplatin-based chemotherapy in advanced colorectal cancer. Clin Cancer Res. 2020 Dec 8:clincanres.3286.2020. doi:10.1158/1078-0432.CCR-20-3286. Epub ahead of print. PMID: 33293373.
Cancer of Unknown Primary (CUP) occurs in 3–5% of patients when standard histological diagnostic tests are unable to determine the origin of metastatic cancer. Typically, a CUP diagnosis is treated empirically and has very poor outcomes, with median overall survival less than one year. Gene expression profiling alone has been used to identify the tissue of origin but struggles with low neoplastic percentage in metastatic sites which is where identification is often most needed.
To assist in providing diagnostic accuracy of CUP patients, Caris Life Sciences® developed a genomic prevalence score called MI GPSai™. This score is an AI-driven tumor type biology similarity score that uses more than 6500 mathematical models in the machine learning algorithm to compare molecular characteristics of a patient’s tumor against Caris’ extensive database to provide new insights into the molecular subtype of cancer of unknown primary (CUP) cases, atypical clinical presentation cases, and other difficult to treat cancer cases, to help guide treatment decisions. It analyzes genomic and transcriptomic data to match a tumor’s molecular signature across 21 cancer types from the Caris database. MI GPSai is intended to provide additional insight to help oncologists better manage CUP or cases with atypical clinical presentation or clinical ambiguity, as identified by the ordering physician.
Caris’ MI GPSai algorithm trained on genomic data from over 34,000 cases and genomic and transcriptomic data from more than 23,000 cases and was validated on over 19,500 cases. MI GPSai predicted the tumor type in the labeled data set with an accuracy of over 94% on 93% of cases while deliberating amongst 21 possible categories of cancer. When also considering the second highest prediction, the accuracy increases to 97%. Additionally, MI GPSai rendered a prediction for 71.7% of CUP cases. The results were published in Translational Oncology1 , a part of Elsevier’s Oncology Journal Network. Previous data shared at the 2019 American Society of Clinical Oncology (ASCO) Annual Meeting showed this score classified tumors from 55,780 samples with over 95% accuracy and generated an unequivocal result in the vast majority of CUP cases, when there was ambiguity about tissue of origin.
MI GPSai provides clinically meaningful information in a large proportion of CUP cases and inclusion of MI GPSai in clinical routine could improve diagnostic fidelity. Moreover, all genomic markers essential for therapy selection are assessed in this assay, maximizing the clinical utility for patients within a single test.
1 Jim Abraham, David Spetzler, et al., Machine learning analysis using 77,044 genomic and transcriptomic profiles to accurately predict tumor type, Translational
Oncology, Volume 14, Issue 3, 2021, 101016, ISSN 1936-5233, (https://go.carislifesciences.com/e/711053/-article-pii-S1936523321000085/grkt3/415450212?h=506mbGp9bruV4FIZtoBveotzINW8zLeQNK4aQyBgIIQ)