Welcome to the CCO Site

Thank you for your interest in CCO content. As a guest, please complete the following information fields. These data help ensure our continued delivery of impactful education. 

Become a member (or login)? Member benefits include accreditation certificates, downloadable slides, and decision support tools.


CCO Independent Conference Highlights of the Virtual 2021 Gastrointestinal (GI) Cancers Symposium

Axel Grothey Headshot
Axel Grothey, MD
Manish A. Shah, MD
Released: March 10, 2021

Colorectal Cancer

KEYNOTE-177: Pembrolizumab vs Chemotherapy for MSI-H/dMMR mCRC

Axel Grothey, MD:
We will begin our discussion with the phase III KEYNOTE‑177 trial.[1-3] KEYNOTE-177 was a randomized, open-label study enrolling 307 treatment-naïve patients with microsatellite instability–high/mismatch repair–deficient (MSI-H/dMMR) metastatic colorectal cancer (mCRC) based on local testing. Enrolled patients had an Eastern Cooperative Oncology Group (ECOG) performance status (PS) of 0 or 1 and measurable disease. Randomization was 1:1 to treatment with pembrolizumab (n = 153) or investigator's choice of doublet chemotherapy (n = 154). Dual primary endpoints were PFS and OS. Crossover was permitted at disease progression.

Patients in the first cohort received 200 mg of pembrolizumab every 3 weeks for up to 35 cycles. Patients in the second cohort group received modified leucovorin/fluorouracil/oxaliplatin (mFOLFOX6) or leucovorin/fluorouracil/irinotecan (FOLFIRI) with or without bevacizumab or cetuximab. Patients continued therapy until progressive disease, intolerable toxicity, or patient withdrawal.

The analysis presented by Shiu and colleagues[2] at ASCO GI 2021 is an updated analysis evaluating PFS after next line of therapy or death from any cause (PFS2) and health-related quality of life (QoL). Although the OS data remain immature, the final PFS analysis demonstrated superiority of pembrolizumab over chemotherapy.

KEYNOTE-177: Baseline Characteristics

Axel Grothey, MD:
KEYNOTE-177 was the first randomized trial evaluating first-line pembrolizumab in patients with mCRC. This was critical for non‑US sites, especially in Europe, because pembrolizumab was not approved there for the treatment of patients with MSI‑H cancers.

Baseline characteristics were well balanced between the arms. The median patient age was about 63 years. Half of the patients were male, had an ECOG PS of 0, and had metachronous disease. Of note, about 25% of patients had BRAF V600E–mutant tumors. A major question is how patients with BRAF-mutant tumors respond to first-line pembrolizumab treatment compared with non–BRAF-mutant tumors.


Axel Grothey, MD:
The new data presented at ASCO GI 2021 looked from randomization to PFS2, taking crossover into account. Median PFS2 was not reached in the pembrolizumab group and was 23.5 months in the chemotherapy group (HR: 0.63; 95% CI: 0.45-0.88). The 12-month rate of PFS2 was 76% for pembrolizumab and 67% for chemotherapy, and the 24-month PFS2 was 65% and 50%, respectively. There was a statistically significant difference in favor of pembrolizumab-first administration, and 36% of patients crossed over (59% in the intent-to-treat population), showing why the analysis was important. Although the OS data are not mature, they will likely be presented this year at ASCO.

KEYNOTE-177: Changes in EORTC QLQ-C30 Scores From Baseline to Week 18

Axel Grothey, MD:
The European Organization for Research and Treatment of Cancer QoL questionnaire (EORTC QLQ-C30) was used to evaluate changes in patient symptoms from baseline to Week 18. QoL scores did not decrease over time with pembrolizumab, and there was a trend toward QoL and symptom benefit compared with chemotherapy. Emotional functioning improved in patients in both groups, but social functioning improved much more in the pembrolizumab group. There was a trend toward improvement in global health status with pembrolizumab, but global health status declined with chemotherapy. Pain and appetite loss also improved with pembrolizumab vs chemotherapy. There was a clear benefit in both PFS and QoL for the use of pembrolizumab over chemotherapy.

KEYNOTE-177: Time to Deterioration in Social Functioning and Fatigue per EORTC QLQ-C30

Axel Grothey, MD:
The time to deterioration in social functioning and fatigue was longer for pembrolizumab compared with chemotherapy. For social functioning, the median time to deterioration was not reached in either group (HR: 0.53; 95% CI: 0.32-0.87; P = .0050). For fatigue, a median time to deterioration was not reached for the pembrolizumab arm and was 2.1 months with chemotherapy (HR: 0.48; 95% CI: 0.33-0.69; P < .0001).

KEYNOTE-177: AEs in All Treated Patients

Axel Grothey, MD:
The rate of adverse events (AEs) and treatment-related AEs was similar in both treatment arms. Grade ≥ 3 treatment-related AEs were much more common in the chemotherapy group, affecting 66% of the patients compared with 22% of the patients receiving pembrolizumab. Grade ≥ 3 diarrhea, fatigue, and neutropenia also were significantly higher in the chemotherapy group. These results were expected, as pembrolizumab is generally well-tolerated.

KEYNOTE-177: Immune-Mediated AEs and Infusion Reactions

Axel Grothey, MD:
The rate of immune‑related AEs and infusion reactions with pembrolizumab was low, which would not be the case when giving an anti–CTLA-4 antibody to these patients. I think pembrolizumab is a new standard of care based on the benefits in efficacy and QoL, as well as an acceptable toxicity profile. The FDA made the right decision in approving pembrolizumab for upfront therapy in patients with MSI‑H colorectal cancer (CRC).

KEYNOTE-177: Clinical Implications

Axel Grothey, MD:
In KEYNOTE-177, pembrolizumab significantly improved PFS and durable responses compared with chemotherapy for patients with MSI-H/dMMR mCRC (24-month PFS: 48.3% vs 18.6%). This is proof that we should use anti–PD-1 antibodies in the first‑line setting. I acknowledge that some patients have progressive disease as best response with pembrolizumab compared with chemotherapy, but we cannot yet identify these patients upfront. Dr. Shah, where are we going with that?

Manish A. Shah, MD:
I think you make an excellent point that some patients may benefit more from chemotherapy than upfront pembrolizumab. However, the data show superior long‑term survival and reduced toxicity for pembrolizumab compared with chemotherapy.

One question I have is from a subgroup analysis suggesting lesser benefit for pembrolizumab in KRAS‑mutant tumors because the HR crosses 1. I am not yet prepared to avoid immunotherapy for KRAS‑mutant tumors in the first‑line setting. What do you think?

Axel Grothey, MD:
I completely agree. First of all, the HR crossing 1 does not mean it is detrimental to use pembrolizumab. We are not making a mistake in giving pembrolizumab to patients with KRAS‑mutant tumors. These findings were unexpected, because previous studies did not observe a difference in response based on KRAS/RAS‑mutant status.

Manish A. Shah, MD:
There is at least some rationale, because KRAS is thought to create a more immunosuppressive environment. So it is feasible that there is less benefit. But I think you make a good point that pembrolizumab is not better or worse than chemotherapy.

Axel Grothey, MD:
A key issue that intrigues us all: Do patients with durable responses have a curative chance or a long‑term chance of no active disease? I believe they might, but the future will tell whether we really move some of these stage IV patients into cure.

TASCO1: First-line Trifluridine/Tipiracil Plus Bevacizumab vs Capecitabine Plus Bevacizumab in Unresectable mCRC

Axel Grothey, MD:
Some patients with mCRC are not candidates for aggressive combination therapies such as oxaliplatin and irinotecan. In the first-line setting, these patients are conventionally treated with a fluoropyrimidine—commonly capecitabine—plus bevacizumab. TASCO1 was an exploratory, open-label, noncomparative phase II study of patients with unresectable mCRC naive to systemic therapy and not eligible for intensive therapy (N = 153).[4-6] Patients received trifluridine/tipiracil plus bevacizumab (TT-B; n = 77) or capecitabine plus bevacizumab (C-B; n = 76).

TT-B was received orally at 35 mg/m2 twice daily on Days 1-5 and Days 8-12 of 4-week cycles. Capecitabine was received orally at 1250 or 1000 mg/m2 on Days 1-14 of 3-week cycles. All patients received 7.5 mg/kg IV bevacizumab on Day 1 of 3-week cycles. Patients continued treatment until progressive disease, intolerable toxicity, or investigator/patient decision. The primary endpoint was PFS, and secondary endpoints were OS, ORR, disease control rate, QoL, and safety. Van Cutsem and colleagues[5] reported the current analysis of OS from the end of the study at ASCO GI 2021.

TASCO1: Baseline Patient Characteristics

Axel Grothey, MD:
Enrolled patients were older than what we typically see in clinical trials (median: 74 years of age), because these patients were not considered candidates for oxaliplatin and irinotecan. About 15% of the patients in each group had a baseline ECOG PS of 2. Almost 60% of the patients in both groups had mutant RAS, and 10% had mutant BRAF. The most common reasons enrolled patients were considered ineligible for intensive chemotherapy were age, ECOG PS, and low tumor burden. This is because patients with a low tumor burden do not need combination chemotherapy with oxaliplatin/irinotecan.

TASCO1: End-of-Study OS

Axel Grothey, MD:
In this noncomparative analysis, median end-of-study OS was longer for the combination of TT-B vs C-B: 22.31 months (95% CI: 18.00-23.69) vs 17.67 months (95% CI: 12.58-19.81), respectively. However, when you look at the survival probabilities at different time points, the difference was not large. By 24 months, the survival probability with TT-B was 0.38 (95% CI: 0.27-0.49) vs 0.34 with C-B (95% CI: 0.24-0.45). Subgroup analysis revealed that the treatment effect on OS was significantly in favor of TT-B for women, patients with BRAF mutations, and patients without surgical resection.

TASCO1: Safety

Axel Grothey, MD:
Patients treated with TT-B had a higher rate of all-grade neutropenia (54.5%) vs patients treated with C-B (7.9%). Hand–foot syndrome was a major AE in patients receiving C-B vs TT-B (52.6% vs 3.9%). These data are consistent previous reports and indicate an acceptable safety profile for TT-B.

TASCO1: Clinical Implications

Axel Grothey, MD:
Patients with unresectable mCRC who received first-line TT-B had a median OS of 22.3 months vs 17.7 months with C-B. These data were consistent with data from a Danish study evaluating TT-B as later-line therapy.[7]

These data are encouraging, but they are from a noncomparative phase II trial. Fortunately, data from the confirmatory phase III SOLSTICE trial are expected later this year.[8]

CCTG CO.26: Tissue and Plasma TMB as Predictive Biomarkers for Immunotherapy Benefit in mCRC Patients

Axel Grothey, MD:
The randomized phase II CCTG CO.26 trial evaluated durvalumab and tremelimumab as a later‑line therapy in patients with microsatellite-stable mCRC.[9,10] Efficacy data were previously reported, and the data presented by Loree and colleagues[9] at ASCO GI 2021 addressed whether tumor mutational burden (TMB) in tissue or plasma was a predictive biomarker for immunotherapy benefit. A circulating tumor DNA (ctDNA) analysis was used to determine TMB.

The TMB within the tissue and plasma was not correlated (r = -0.039; R2 = 0.002; P = .69), and plasma had a significantly higher TMB compared with tissue (median 16.3 muts/Mb vs 6.6 muts/Mb; P < .0001). This may have been the result of a large time gap between the archival tissue and plasma collection (median 3.1 years; interquartile range: 1.9-5.1) or previous treatment increasing plasma TMB.

CCTG CO.26: High Plasma TMB, but Not High Tissue TMB, Is a Negative Prognostic Marker

Axel Grothey, MD:
The high TMB cutoff points of 10 muts/Mb for tissue and 28 muts/Mb for plasma analysis were previously established.[11,12] High tissue TMB was not prognostic of median OS. In patients with TMB ≥ 10 muts/Mb (29%), median OS was 4.47 months compared with 4.27 months in patients with < 10 muts/Mb (HR: 1.01; 90% CI: 0.52-1.92; P = .99). By contrast, high plasma TMB was associated with a decrease in median OS. In patients with plasma TMB ≥ 28 muts/Mb (21%), median OS was 2.96 months compared with 5.26 months for patients with plasma TMB < 28 muts/Mb (HR: 2.56; 90% CI: 1.45-4.54; P = .007). This is a phenomenon that we do not know how to exploit yet, but it suggests that ctDNA obtained immediately before new lines of therapy is more relevant than archival tissue from several lines of therapy earlier.

CCTG CO.26: High Plasma TMB Predictive of Durvalumab Plus Tremelimumab Efficacy

Axel Grothey, MD:
Plasma TMB ≥ 28 muts/Mb was predictive of durvalumab plus tremelimumab efficacy (P = .004), but tissue TMB ≥ 10 muts/Mb was not (P = .14). Tissue TMB < 4.1 muts/Mb may predict activity, but the results are inconclusive and may be from a graphical artifact (P = .017).

CCTG CO.26: Clinical Implications

Axel Grothey, MD:
CCTG CO.26 results suggest that high plasma TMB—but not high tissue TMB—is a negative prognostic marker in patients with mCRC. I think these data are hypothesis generating and need to be confirmed. I would not use durvalumab and tremelimumab or other immunotherapies in patients with microsatellite-stable tumors outside of clinical trials.

Manish A. Shah, MD:
It is interesting that plasma TMB was prognostic and not tissue TMB. We think of blood-based TMB as a more comprehensive analysis of the tumor, but tumor cells in the blood could be more aggressive or metastatic. More work is needed to determine what the circulating tumor cells reflect and why they are more prognostic than tumor cells from the tissue.

Analysis of Relationship Between ctDNA and Recurrence Risk in Colorectal Cancer: Study Design

Axel Grothey, MD:
In CRC, ctDNA is a mark of measurable residual disease after resection of early stage cancer. Many cohort studies and retrospective analyses have looked at the implication of ctDNA positivity in long term patient outcomes and found that patients who are initially ctDNA positive have an almost 100% risk of recurrence.[13] However, it is possible that treatment with adjuvant therapy protects patients who were initially ctDNA positive against recurrence. This study used whole exome sequencing of tumor and peripheral blood mononuclear cells to longitudinally measure ctDNA positivity over 3 years. [14] The study aims were to stratify patients at high risk or low risk of recurrence based on detection of measurable residual disease, to assess posttherapy relapse risk in ctDNA-positive patients, and to determine lead time of ctDNA detection compared with chemotherapy recurrence. The results of this study were presented by Henriksen and colleagues[14] at ASCO GI 2021.

ctDNA and CRC Recurrence: Patient Baseline Characteristics

Axel Grothey, MD:
This study analyzed 260 patients with stage I (n = 4), stage II (n = 90), and stage III (n = 166) CRC. In total, 48 patients relapsed, and 165 were adjuvant treated. The median follow-up for nonrelapsed patients was 29.9 months (range: 1.2-51.0).

Postoperative ctDNA Detection and CRC Recurrence

Axel Grothey, MD:
Patients who were ctDNA positive after resection (n = 20) had a significant reduction in relapse-free survival compared with patients who were ctDNA negative (n = 198; HR: 11; 95% CI: 5.9-21; P < .0001). These findings were consistent with previous data. Within the 29.9-month median follow-up time frame, recurrence occurred in 80% of ctDNA-positive patients and only 13% of ctDNA-negative patients. The ctDNA-positive patients remaining relapse free were likely a factor of adjuvant chemotherapy, because recurrence normally occurs in all ctDNA-positive patients within 3 years.

Posttreatment ctDNA Detection and CRC Recurrence

Axel Grothey, MD:
Patients have a poor prognosis if they become ctDNA positive after adjuvant chemotherapy. Patients who were ctDNA positive after the end of adjuvant chemotherapy (n = 12) had a recurrence rate of 83.3% vs only 12.5% for ctDNA-negative patients (n = 96; HR: 12; 95% CI: 4.9-27; P < .0001). This trend was also observed during longitudinal monitoring, as patients who were ctDNA positive (n = 28) had 89.3% recurrence vs only 3.4% in ctDNA-negative patients (n = 174; HR: 51; 95% CI: 20-125; P < .0001).

Time to CRC Recurrence by CT Scan and ctDNA Assessment

Axel Grothey, MD:
The time from positive ctDNA to detection of recurrence (as measured by CT) was a median of 8.1 months (interquartile range: 2.2-11.4 months). This varies among cancers, but the time between ctDNA positivity and imaging recurrence is relatively long for CRC.

ctDNA a Better Predictor of Recurrence Than CEA

Axel Grothey, MD:
ctDNA was a much better predictor of recurrence than carcinoembryonic antigen expression. Longitudinally and by multivariable analysis, the carcinoembryonic antigen HR for disease recurrence was 1.8 (95% CI: 0.77-4.0; P = .184), and the HR for ctDNA was 80.55 (95% CI: 23.1-281; P < .0001). In addition, about 20% of patients with CRC are negative for carcinoembryonic antigen, so it cannot be used as a marker of recurrence in all patients.

ctDNA and Recurrence Risk in Colorectal Cancer: Clinical Implications

Axel Grothey, MD:
Longitudinal ctDNA was predictive of recurrence risk and outperformed carcinoembryonic antigen testing, so I think it will emerge as a major tool to predict recurrence and influence adjuvant therapy. Several prospective clinical trials are evaluating the use of ctDNA testing to delay adjuvant therapy until ctDNA positivity.[15,16]

Manish A. Shah, MD:
I have had a few patients who were clinically doing well and feeling fine 1 year off treatment who became ctDNA positive. They were devastated because of the high chance of recurrence and did not know what to do. I think future studies will ask whether we can reverse ctDNA positivity. The ctDNA positive patients who did not die might reflect adjuvant therapy; it is unlikely the ctDNA results were false positives. We definitely need more data.

Phase II NRG-GI002 Trial of Total Neoadjuvant Therapy (TNT) in LARC: Pembrolizumab ARM

Axel Grothey, MD:
Rectal adenocarcinomas are a collection of diseases that vary in location, goal of care, and management approaches. The randomized, modular phase II NRG-GI002 trial is evaluating upfront total neoadjuvant therapy in patients with locally advanced rectal adenocarcinoma.[17] Patients were previously untreated and had stage II/III disease with an ECOG PS of 0-2 and adequate organ function. Randomization was 1:1 to pembrolizumab plus capecitabine and mFOLFOX6 (n = 90) or capecitabine and mFOLFOX6 alone (n = 95). This presentation discussed results from the pembrolizumab arm of the trial (N = 185).[18]

During the first 8 cycles of treatment, patients received mFOLFOX6 every 2 weeks. Radiation therapy (RT) and capecitabine were started 3-4 weeks after the last dose of mFOLFOX6, and 825 mg/m2 of capecitabine was received orally twice daily on the days patients received RT. In patients who received pembrolizumab, 200 mg was administered IV every 3 weeks starting on Day 1 of RT for 6 total doses. Surgery was performed 8-12 weeks after the last RT dose. The primary endpoint was neoadjuvant rectal score, a tissue‑based analysis of regression grade based on criteria from previous studies. Secondary endpoints included pathologic CR, composite CR, OS, disease-free survival, and toxicity.

NRG-GI002: Primary and Secondary Endpoints

Axel Grothey, MD:
The mean neoadjuvant rectal score was similar in the pembrolizumab and control groups (11.53 vs 14.08; P = .26). There also were no significant differences between groups for pathologic CR (P = .75), clinical CR (P = .95), margin-negative resection (P = .36), and sphincter preservation (P = .15). Of note, only 46% of patients completed all 6 pembrolizumab doses. An additional 26% completed 5 doses, and the remaining 26% completed 1-4 doses.

NRG-GI002: Safety

Axel Grothey, MD:
No major toxicities were associated with pembrolizumab treatment. Pembrolizumab related AEs such as hypothyroidism occurred, but overall it was well tolerated.

NRG-GI002: Clinical Implications

Axel Grothey, MD:
The addition of pembrolizumab to mFOLFOX6 and capecitabine did not improve neoadjuvant rectal score or any of the reported secondary endpoints compared with mFOLFOX6 and capecitabine alone. The use of pembrolizumab in patients with locally advanced rectal adenocarcinoma was not supported.

Despite these results, I believe we are moving toward total neoadjuvant therapy for the majority of patients with rectal cancer, with the sequence of events being RT followed by chemotherapy. What kind of RT? Short course vs long course? What kind of chemotherapy? Those questions are still being ironed out, but I think total neoadjuvant therapy will become the standard of care. It may also increase the likelihood of patients receiving nonoperative management, which is clinically meaningful.

Manish A. Shah, MD:
The watch-and-wait approach for rectal cancer is an interesting paradigm, and I wonder if it will ever be compared with a surgical approach. While we are waiting for the results of the PROSPECT study, which asks whether RT is needed,[19] I feel uncomfortable with chemotherapy alone. I want the radiation to ensure that microscopic disease in the rectal bed is killed.

Axel Grothey, MD:
I agree. If we want to avoid surgery, we should maximize everything we do locally, including RT. Radiation de‑escalation is already happening in studies comparing short course and long course RT. PROSPECT is evaluating the selective use of RT. A phase II study of nonoperative management is evaluating whether surgery is necessary.[20] The only component that I think is here to stay is chemotherapy.

Provided by Clinical Care Options, LLC

Contact Clinical Care Options

For customer support please email: customersupport@cealliance.com

Mailing Address
Clinical Care Options, LLC
12001 Sunrise Valley Drive
Suite 300
Reston, VA 20191

Supported by educational grants from
Daiichi Sankyo, Inc.
Merck Sharp & Dohme Corp.
Seagen Inc.
Taiho Oncology Inc.

Leaving the CCO site

You are now leaving the CCO site. The new destination site may have different terms of use and privacy policy.


Cookie Settings