2026 ASCO Annual Meeting

GONO studies presented during 2026 ASCO Annual Meeting:

• Molecular hyperselection in proficient mismatch repair/microsatellite stable (pMMR/MSS) RAS and BRAF wild-type metastatic colorectal cancer – By Dr. Marco Maria Germani
• FOLFOXIRI/bevacizumab in microsatellite stable (pMMR/MSS) RAS and BRAF wild-type, “gene altered” metastatic colorectal cancer (mCRC): the HYPERTRIBE study – By Dr. Marco Maria Germani
• Plasma-TMB and circulating tumor fraction dynamics in the AtezoTRIBE study – By Dr. Vittorio Studiale
• Development of a computational histology artificial intelligence (CHAI)–powered predictive biomarker for first-line chemotherapy intensification in metastatic colorectal cancer and validation in prospective randomized phase III trials – By Dr. Paolo Ciracì
• A transcriptomic signature predicts immune checkpoint inhibition efficacy in proficient mismatch repair metastatic colorectal cancer: Translational analysis of the AVETRIC and TRIPLETE studies – By Dr. Martina Carullo

At ASCO Annual Meeting 2026, Dr. Marco Maria Germani, on behalf of GONO Foundation, presented two posters related to first-line treatment personalization in proficient mismatch pMMR/MSS RAS/BRAF wild-type mCRC. According to current international guidelines, first-line treatment is represented by a doublet plus anti-EGFRs in left-sided and doublet or triplet (in clinically fit patients) plus the antiangiogenic bevacizumab (bev) in right-sided tumours belonging to this molecular subgroup. The underlying biological rationale is represented by unresponsiveness to EGFR inhibition in right-sided mCRC, that is enriched of genomic alterations beyond RAS/BRAF mutations driving resistance to anti-EGFRs (i.e. MET amplification, ALK/ROS1/NTRKs/RET fusions, and HER2/PIK3CAexon20/PTEN/AKT1 mutations). These alterations can be detected in left-sided tumour, as well. In other words, primary tumor location may not be most accurate predictive biomarker of choice between anti-EGFRs or bev and comprehensive genomic profiling (CGP) may help to fine-tune the use of monoclonal antibodies in pMMR/MSS RAS/BRAF wild-type mCRC, regardless of primary tumor location.

Negative hyperselection as predictive biomarker of benefit from first-line doublets + anti-EGFRs/bevacizumab (bev) in microsatellite stable (pMMR/MSS) RAS and BRAF wild-type and HER2- metastatic colorectal cancer (mCRC): an individual patient data pooled analysis of seven clinical trials (CTs)

In the first poster, we collected individual patient data from 1198 pMMR/MSS RAS/BRAF wild-type mCRC with CGP available, enrolled in seven academic phase II/III clinical trials in Italy, Germany, US and Japan. HER2-positive cases were excluded, because last year at ASCO we demonstrated that HER2 expression is a prognostic, rather than a predictive biomarker of benefit from first line combinatory chemotherapy plus anti-EGFRs/bev. All eligible patients were treated with doublets plus anti-EGFRs or bev. Left and right-sided tumors were 990 (83%) and 208 (17%), respectively. 121 (10%) harboured at least one genomic alteration associated with resistance to anti-EGFRs beyond RAS/BRAF (“gene altered” cohort). The remaining 1077 (90%) were devoid of any of these alterations (“hyperselected” cohort). As expected, anti-EGFRs were associated with an overall survival (OS) benefit versus bev in left-sided (HR: 0.87, 95% CI: 0.75 – 1.01, p=0.065) but not right-sided tumors (HR: 0.97, 95% CI: 0.72 – 1.31, p=0.859) with a p-value for interaction of 0.539. Importantly, when the patients were stratified according to molecular profile, anti-EGFRs were associated with an overall survival (OS) benefit versus bev in hyperselected (HR: 0.84, 95% CI: 0.73 – 0.97, p=0.017) but not in gene-altered tumors (HR: 1.17, 95% CI: 0.76 – 1.76, p=0.495), with a p-value for interaction dropping to 0.113, suggesting that molecular profiling may be a more accurate biomarker of benefit from anti-EGFRs or bev in pMMR/MSS RAS/BRAF wild-type HER2-negative mCRC patients in their first line of therapy, as compared to primary tumor location. From a practical standpoint, the most relevant finding of this work is that right-sided and hyperselected tumors may be eligible for an anti-EGFR containing first-line doublet regimen, while some left-sided and gene-altered tumors yield similar survival outcomes with anti-EGFR and bev.

In the second poster, we focused exclusively on the “gene altered” cohort, where we investigated in the observational study called “HYPERTRIBE” whether an intensified chemotherapy regimen with the triplet FOLFOXIRI plus bev could represent a valuable option, given that both doublets plus bev and doublets plus anti-EGFR are associated with similar poor efficacy outcomes. Overall, 35 pMMR/MSS RAS/BRAF wild-type “gene altered” patients were enrolled. With a median progression-free survival (PFS) of 12.3 months and the lower boundary of the 80% confidence interval exceeding 9 months, corresponding to the expected median PFS from doublets plus anti-EGFRs (80% CI: 9.2 – 14.3), the primary endpoint was met (Brookmeyer-Crowly test: 0.023). Median OS and objective response rate (ORR) were 35.9 months (95% CI: 28.3 – 55.5) and 74%, respectively. Meaningfully, when compared with an external propensity-score matched cohort of 35 pMMR/MSS RAS/BRAF wild-type “gene altered” patients treated with doublets plus anti-EGFRs, FOLFOXIRI plus bev demonstrated a numerical advantage in terms of PFS (HR: 0.73, 95% CI: 0.45 – 1.21, p=0.226), OS (HR: 0.57, 95% CI: 0.32 – 1.03, p=0.058), and ORR (OR: 1.69, 95% CI: 0.61 – 4.88, p=0.303). Taken together, these data suggest that FOLFOXIRI plus bev retains efficacy also in pMMR/MSS RAS/BRAF wild type “gene altered” patients, with efficacy outcomes consistent with historical data of the triplet backbone. Also, we demonstrated that FOLFOXIRI plus bev may potentially outperform doublets plus anti-EGFRs in this molecular subgroup.

At the 2026 ASCO Annual Meeting held in Chicago, US, Dr. Vittorio Studiale presented the results of a collaborative project between GONO GI investigators and Tempus AI (Chicago, US). This retrospective analysis investigated the dynamics of two plasma biomarkers, plasma tumor mutational burden (pTMB) and circulating tumor fraction (cTF), as potential predictors of benefit from the addition of atezolizumab to first-line FOLFOXIRI + bevacizumab in metastatic colorectal cancer (mCRC).

In the phase II AtezoTRIBE trial, tissue TMB (tTMB) emerged as a promising predictive biomarker of benefit for the experimental arm, despite no ORR improvement and PFS curves separating only after approximately 6 months. We thus hypothesized that dynamic liquid biopsy biomarkers could better capture early molecular changes linked to subsequent immune sensitivity and improved outcomes.

Plasma samples of patients included in the trial were collected at baseline (T0) and at the end of induction treatment (T1). Samples were analyzed using the Tempus xF assay, a 105-gene cell-free DNA liquid biopsy panel, to determine the baseline pTMB and cTF. For the pTMB analysis, a cut-off of ≥26 mut/Mb was used to stratify patients; for cTF dynamics, patients with the top 30% relative cTF reduction between T0 and T1 were defined as molecular responders (MRs).

In the overall cohort, baseline high pTMB showed a trend towards longer PFS (HR: 0.69 [90% CI 0.43-1.10], p = 0.095). However, in a sensitivity analysis excluding 8 dMMR/MSI-H patients, the observed prognostic signal was attenuated, while remaining in the expected direction (HR: 0.74 [90% CI 0.45-1.22], p = 0.163). Notably, achieving a molecular response in terms of cTF reduction was associated with prolonged PFS compared to non-MRs. Moreover, in the full cohort, cTF dynamics showed a potential predictive value, with non-MRs achieving greater benefit from the addition of atezolizumab compared to MRs (pint = 0.132). Nonetheless, this predictive value was not confirmed in the sensitivity analysis excluding 6 dMMR/MSI-H patients (pint = 0.462).

These data suggest that baseline high pTMB may hold a potential prognostic trend, and molecular response to first-line FOLFOXIRI + bevacizumab +/- atezolizumab measured by cTF dynamics is associated with longer PFS in pMMR mCRC patients. However, the predictive utility of both plasma biomarkers for immunotherapy benefit appears largely driven by the dMMR/MSI-H subpopulation, highlighting the need for improved biomarker-driven selection in pMMR mCRC cohorts.

At the 2026 ASCO Annual Meeting in Chicago, Dr. Paolo Ciracì of the GONO GI Foundation presented the development and validation of a novel artificial intelligence (AI)-powered computational histology biomarker designed to predict the benefit of first-line chemotherapy intensification in patients with metastatic colorectal cancer.

The biomarker was developed in collaboration with VALAR Labs (Palo Alto, California) using the Computational Histology Artificial Intelligence (CHAI) platform, a pan-cancer foundation model that analyzes routine hematoxylin and eosin (H&E)-stained pathology slides to capture key histomorphological features and generate predictive biomarkers to support the therapeutic decision-making process.

The AI model was trained using whole-slide images (WSIs) from patients enrolled in the randomized phase III FOCUS trial. After the optimal threshold was established, the biomarker was locked and independently validated using WSIs from patients enrolled in the randomized phase III TRIBE and TRIBE-2 trials by the GONO Foundation.

In the validation cohort, among biomarker-positive patients (48%), treatment with FOLFOXIRI plus bevacizumab significantly improved both progression-free survival (14.7 versus 8.6 months) and overall survival (30.1 versus 19.9 months) compared with doublet chemotherapy plus bevacizumab. In contrast, biomarker-negative patients (52%) derived no significant benefit from treatment intensification. Notably, the AI-powered biomarker also retained its predictive value in the clinically relevant subgroup of patients with right-sided and/or RAS-mutated disease.

These findings highlight the potential of the mCRC CHAI biomarker to support a more personalized first-line treatment selection for patients with metastatic colorectal cancer. By leveraging routine H&E-stained slides, the biomarker may help identify patients most likely to benefit from upfront intensified chemotherapy regimens, thereby optimizing survival outcomes while minimizing unnecessary treatment-related toxicity.

At ASCO Annual Meeting 2026, Dr. Martina Carullo, on behalf of GONO GI Foundation, presented results of a transcriptomic signature designed to predict benefit from immune checkpoint inhibitors (ICIs) in proficient mismatch repair (pMMR) metastatic colorectal cancer (mCRC) patients. ICIs tipically provide restricted benefit in pMMR mCRC, highlighting the need for predictive biomarkers.

We aimed to uncover transcriptomic profiles by selecting two matched cohorts from two academic clinical trials involving a molecularly selected population: RAS and BRAF wild-type (wt) pMMR mCRC who are typically treated with chemotherapy combined with anti-EGFR agents (panitumumab or cetuximab) as first-line therapy. Specifically, the AVETRIC trial provided the ICI-treated cohort (mFOLFOXIRI + cetuximab + avelumab) while the experimental arm of the TRIPLETE trial (mFOLFOXIRI + panitumumab) served as the ICI-untreated control cohort.

Bulk RNA sequencing analysis [Signomax™ platform] were performed on naïve tumour specimens collected from both cohorts. Particularly, gene set variation analyses compared extended progression-free survival (e-PFS) versus limited PFS (l-PFS) groups, leading to focus on immune- and EGFR related pathways. Cases with both pathways classified as high were combined into the immunehigh-EGFRhigh signature and compared with a pooled group comprising all remaining cases (immune-EGFRothers). Thirty-four (87%) and 26 (81%) samples were successfully analyzed in the ICI and control cohorts, respectively. Expression levels of both immune- and EGFR-related pathways were higher in patients with e-PFS compared to the l-PFS group within the AVETRIC cohort. In contrast, no significant differences in pathway expression were observed between PFS groups in the ICI-untreated control cohort. Additionally, 29 cases were classified as immunehigh-EGFRhigh and 31 as immune-EGFRothers. Notably, immunehigh-EGFRhigh signature was associated with significantly longer PFS (HR: 0.57, 95% CI 0.33-0.98, P= 0.04) and overall survival (OS) (HR : 0.66, 95% CI 0.36-1.24, P=0.20) compared to the Immune-EGFRothers group. Furthermore, the benefit derived from the addition of the ICI avelumab to chemotherapy with mFOLFOXIRI and anti-EGFRs appeared to be restricted to patients bearing the immunehigh-EGFRhigh signature (HR PFS : 0.56, 95% CI .25-1.24, versus 1.25, 95% CI 0.60-2.64 in the immune-EGFRothers; Pinteraction= 0.145).

These findings suggest that transcriptomic profiling may identify a subset of pMMR RAS/BRAF wt mCRC with concurrent immune activation and EGFR dependency that could benefit from combining ICIs to an intensified anti-EGFR-based first-line treatment.