GI Malignancies - dMMR (June 2019)

ASCO University
Jun 12, 2019 11:18 AM

Participant Instructions: Welcome to the Multidisiplinary Molecular Oncology Tumor Board Series! This educational initiative is a collaboration between the American Society of Clinical Oncology (ASCO), College of American Pathologists (CAP), and Association for Molecular Pathology (AMP).

A new case will be presented bi-monthly with discussions led by an expert pathologist and medical oncologist. This month’s topic is led by Drs. Timothy Cannon (Hematologist/Oncologist from INOVA Medical Group) and Chanjuan Shi (Pathologist from Vanderbilt).

Do you have an interesting case in mind? Submit your hypothetical patient cases for consideration in an upcoming Multidisciplinary Molecular Tumor Board discussion forum.

Participants are encouraged to leave comments and post questions about the case in order to generate a wide discussion among the cancer care community. You can also receive email notifications when new comments are posted by clicking the “Follow this Conversation” option located at the bottom of this page.

When posting, please abide by the terms and conditions of this website.



ASCO University
Re: GI Malignancies - dMMR (June 2019)
Jun 12, 2019 11:22 AM


Patient Case1: 

Mr. Fox is a 68-year-old gentleman with no significant past medical history who presented with worsening nausea and vomiting in late 2016. In October, he noted mild nausea and some dysphagia to solid food. In early December, he began to experience consistent post-prandial vomiting.  He was admitted to a local hospital and an upper endoscopy showed a malignant appearing duodenal stricture. The biopsy showed a moderately differentiated adenocarcinoma. A CT scan the day following the upper endoscopy showed duodenal wall thickening and three metastatic appearing lesions in the liver, in segments 4A, 6, and 7. The largest of the three lesions was in segment 4A, measuring approximately 4.2 cm x 3.8 cm


He received four cycles of 5-fluorouracil, leucovorin and oxaliplatin (modified FOLFOX6). He tolerated this poorly, with grade 2 neuropathy, grade 2 fatigue, and significant frustration with fatigue and the logistics of treatment. His CT scan after cycle 4 demonstrated a partial response based on improvement in his liver lesions. Unfortunately, he began to have increasing difficulty with dysphagia, and was admitted to the hospital with a suspected bowel obstruction. In March 2017, he developed small bowel obstruction and had an exploratory surgery with a gastrojejunostomy (GJ) bypass. He began eating well with no limitations. He refused further FOLFOX.


ASCO University
Re: GI Malignancies - dMMR (June 2019)
Jun 12, 2019 11:24 AM


Discussion Questions:

  1. Which tests are needed to determine whether he has Lynch syndrome?
  2. Which test(s) can determine whether he is a candidate for anti-PD-1 therapy?
  3. What other therapies are potentially available to him?
Saleem Ahmad Siddiqui, MD, FRCPC
Re: GI Malignancies - dMMR (June 2019)
Jun 12, 2019 2:02 PM

  • DNA mismatch repair genes

PDL1        Ramucirumab +Paclitaxel

Lin Rongbo, MD
Re: GI Malignancies - dMMR (June 2019)
Jun 13, 2019 7:58 PM

  1. MMR 
  3. paclitaxel+5-fu/lv+oxaliplatin+apatinib, first, should improve the condition of the patient, response rate is vital importance. although the patient have already got the oxaliplatin, the combination can enhance the effect of other agents. apatinib can synergize with other agents. after his condition improves, consider the immunotherapy according to testing
Ayesha Saqib, MBBS, FRACP
Re: GI Malignancies - dMMR (June 2019)
Jun 17, 2019 6:41 AM

1-Colonoscopy + MMR 


3- irinotecan based tx, PD1 inhibitors (depending biomarkers) and potential liver directed tx given reluctance and logistics related to systemic, 


Arunangshu Das, FCPS, MBBS
Re: GI Malignancies - dMMR (June 2019)
Jun 12, 2019 1:57 PM

Metastatic Colorealc cancer patients needs upfront evaluation with MSI/dMMR, BRAF and RAS


Clifford A. Hudis, MD, FACP, FASCO
Re: GI Malignancies - dMMR (June 2019)
Jun 12, 2019 2:00 PM

1. Test for mismatch repair gene mutations.

2. MSI (high).

3. Taxane-based palliative therapy?

Saleem Ahmad Siddiqui, MD, FRCPC
Re: GI Malignancies - dMMR (June 2019)
Jun 12, 2019 2:07 PM


Saleem Ahmad Siddiqui, MD, FRCPC
Re: GI Malignancies - dMMR (June 2019)
Jun 12, 2019 2:11 PM



ASCO University
Re: GI Malignancies - dMMR (June 2019)
Jun 17, 2019 10:53 AM


Faculty response on behalf of Dr. Shi (Pathology)

Checkpoint inhibitors have been proved to be very successful in treating some colorectal cancers. Mismatch-repair deficiency (dMMR) and microsatellite instability-high (MSI-H) are effective biomarkers predictive of immunotherapy response in these tumors. Tumors with dMMR are associated with increased tumor mutational burden (TMB). A high mutational burden may give rise to tumor-specific antigens (neoantigens), thereby increasing the possibility of the immune system to recognize and destroy tumor cells in response to checkpoint inhibitor therapies.

It is well known that similar mechanisms underlie small and large intestinal carcinogenesis. The NCCN guidelines recommend treating small intestinal adenocarcinoma in a similar way as colorectal cancers. Therefore, MSI status is often evaluated in patients with small intestinal carcinomas (including duodenal carcinomas). Immunohistochemistry (IHC) for the four MMR proteins (MLH1, MSH2, MSH6 and PMS2) and PCR-based MSI testing are used to evaluate MSI status, with MMR IHC more commonly used in the recent years. In addition, MSI status can be analyzed by next generation sequencing (NGS), which is easily added to cancer-specific NGS panels (i.e. FoundationOne CDx).

Approximately 90-95% of dMMR cancers can be identified by IHC, which show loss of nuclear staining for one or more of the MMR proteins. MLH1, MSH2, MSH6 and PMS2 proteins play a critical role in mismatch recognition and initiation of repair. MLH1 and PMS2 form hMutLα heterodimer; loss of MLH1 invariably results in the degradation of PMS2. Similarly, MSH2 and MSH6 form hMutSalpha heterodimer loss of MSH2 is consistently companied by loss of MSH6. However, the converse is not true because loss of PMS2 or MSH6 does not always cause the degradation of MLH1 or MSH2, respectively. Therefore, when a tumor shows loss of MLH1 and PMS2 or loss of MSH2 and MSH6, it is likely due to defective MLH1 or MSH2, whereas isolated loss of PMS2 or MSH6 indicates a defect in PMS2 or MSH6, respectively. The false-negative results are likely due to MMR gene inactivating mutations (loss of function) with preservation of epitope that can still be detected by IHC.  The major advantage of the immunohistochemical assay is identification of possible defective MMR protein(s).

PCR-based MSI testing is performed to detect the genetic changes associated with dMMR, namely MSI. Approximately 90% of MSI-high tumors can be detected by MSI testing. Currently, most laboratories in the US use a fluorescence-based PCR assay from Promega. The assay includes five mononucleotide repeat (MSI) markers (BAT-25, BAT-26, NR-21, NR-24 and MONO-27) and two pentanucleotide repeat (identity) markers (Penta C and Penta D). Both normal and tumor tissue are analyzed. The results are interpreted as MSI-high when two or more mononucleotide markers show instability, MSI-low when only one marker is unstable, and microsatellite stable (MSS) when all markers are stable.


Approximately 15% of colorectal cancers are MSI-H. It is believed that there are more MSI-H tumors in the small intestine than in the colorectum. MSI-H tumors can be sporadic or hereditary. Most MSI-H tumors are sporadic, caused by hypermethylation of MLH1 promotor, and display loss of MLH1 and PMS2 expression by IHC. Approximately 2-4% of colorectal cancers are Lynch syndrome-associated, MSI-H tumors. Nowadays most laboratories universally screen for Lynch syndrome among patients with newly diagnosed colorectal and endometrial cancers. Some laboratories also screen patients with small bowel carcinomas. When an MSI-H tumor is identified by MMR IHC or PCR-based MSI-testing, additional tests, such as BRAF V600E mutational analysis, MLH1 methylation assay and germline mutational analysis, always follow to determine if the tumor is sporadic or hereditary. When there   is loss of MLH1 and PMS2 expression with MLH1 promotor hypermethylation and no loss of any other proteins (MSH2, MSH6, or PMS2), the tumor is considered sporadic. Otherwise, germline evaluation should be carried out at least for the genes corresponding to the absent proteins on IHC stains to identify Lynch syndrome patients (with germline mutations in the MMR genes). For colorectal cancers with loss of MSH2, a test for EPCAM deletions needs to be included, given that an EpCAM deletion in the 3’ region causes somatic hypermethylation of MSH2.

Universal screening for Lynch syndrome leads to identification of many tumors with dMMR/MSI-H, but with no evidence of MLH1 hypermethylation and no germline mutations. Tumor sequencing of gastrointestinal and endometrial cancers revealed that majority of these tumors had double somatic mutations in the MMR gene, corresponding with the absent protein, and some had one somatic mutation in an MMR gene, with possible loss of heterozygosity. Somatic mutations in these tumors are most commonly found in MLH1 gene, followed by MSH2 gene. A study on MSI-H colorectal cancers showed that there were no differences histologically between tumors with double somatic MMR mutations and those caused by Lynch syndrome. Somatic mutations in MMR genes are considered to be the third mechanism underlying MSI-H carcinogenesis and are a frequent cause of dMMR in gastrointestinal cancers.



ASCO University
Re: GI Malignancies - dMMR (June 2019)
Jun 17, 2019 10:56 AM


Faculty response on behalf of Dr. Cannon (Medical Oncology)

Lynch syndrome, by definition, is a germline mutation in mismatch repair genes as described by Dr. Shi. As an aside, I was saddened to hear about the recent passing of Henry Lynch, who is credited with identifying this extremely important phenomenon. At our center, as in most US centers, IHC staining for absence of MM genes is done reflexively for all colorectal adenocarcinomas.  Given that some retrospective studies have shown a similar incidence of dMMR in duodenal cancer, we also reflexively test for MMR on tumor tissue in these situations. This case is fascinating in that it appears to be an example of a somatic mutation causing dMMR, either through biallelic mutation or one mutation with LOH.

As for the second question, currently only dMMR can lead you to an indication for an immune checkpoint inhibitor off trial. However, tumor mutational burden appears to be predictive of a higher likelihood of response in MSS colorectal cancer. This past ASCO, Chen et al presented data for CCTG CO.26, which compared Durvalumab and Tremelimumab versus Best supportive care. For tumor mutational burden above 28 mutations/megabase, the OS benefit had a HR of 0.34. It is worth noting that the pre-specified cut-off was 20 and they used a minimum p=value approach to arrive at a cutoff of 28. At the very least, a search for immunotherapy clinical trials should be considered for high TMB MSS patients. Of course, PD-L1 expression is predictive in many cancers as well.

As far as the third question, he would be eligible for FOLFIRI and Bevacizumab, though this option was not attractive to him. We would also have considered him for TAPUR or NCI Match, given his PIK3CA mutation. It is noteworthy that PBRM1 inactivating mutations have been associated with a higher likelihood of response to immune checkpoint inhibitors.


ASCO University
Re: GI Malignancies - dMMR (June 2019)
Jun 17, 2019 12:22 PM


Patient Case 1 Update

The patient’s initial biopsy in December yielded insufficient tissue for IHC staining for mismatch repair proteins. However, the IHC on the tumor tissue from the March surgery demonstrated loss of MLH1 and PMS2. MLH1 methylation was not detected. Genomic sequencing showed an MLH1 truncation at R226, a PBRM1 frameshift (R906fs*12), and an ERBB3 V104M mutation. There were also mutations in PIK3CA, RAF1, FANCA, LZTR1, and inactivating SMAD4 mutations. The tumor mutational burden was 28 mutations/mb. Germline testing did not detect the MLH1 mutation.

Mr. Fox started on pembrolizumab. After four cycles of treatment, he had less SD with disease shrinkage. After 10 cycles of pembrolizumab, his largest liver lesion measured 1.6 cm x 1.3 cm. As of April 2019, his segment 4A lesion measures 17 mm x 8 mm as seen in below figure. He has continued on pembrolizumab up to the time of writing (May of 2019).



ASCO University
Re: GI Malignancies - dMMR (June 2019)
Jun 17, 2019 12:29 PM


Patient Case 2

Mr. Scott is a 68-year-old gentleman with a history of prostate cancer who presented with stage IV distal esophageal adenocarcinoma.

On June 13, 2017, he had a PET scan that showed an intensely FDG avid, bulky distal esophageal mass, with a maximum SUV of 19.  There was bulky retroperitoneal and peripancreatic adenopathy. There were multifocal liver lesions throughout the left lobe, caudate lobe, and inferior tip of the right lobe.  The most hyper metabolic left- lobe lesion had an SUV of 15.3.

He started chemotherapy with FOLFOX in July.  He received a total of four cycles in Michigan and tolerated it quite well.  He noticed some improvement in his dysphagia.  However, he noticed that the improvement was mostly early in the cycle, with some regression later in the cycle.  He was also noting significant weight loss throughout the chemotherapy.

Mr. Scott moved to our region from Michigan in August of 2017 restarted FOLFOX here shortly after.  His CT scan is below. 

He developed worsening dysphagia during cycle 5 and had placement of a stent. Shortly thereafter, he went to the emergency department for fever. After cycle 6, on September 25, he had a PET scan which showed: decrease in size and SUV of primary mass and liver lesions, but significant growth in a peripancreatic lymph node (3.8 cm x 3.6 cm, previously 2 cm x 2 cm). Most other retroperitoneal nodes have improved.

An endoscopic ultrasound (EUS) biopsy of the RP node was performed. During the procedure, it was discovered that the stent had migrated. It was taken out. The obstruction was fairly mild. He was started briefly on paclitaxel/ramucirumab but tolerated this poorly. Genomic sequencing was ordered. He was found to have microsatellite instability, with a tumor mutational burden (TMB) of 41 mutations per megabase. There was no mutation found in MLH1, MSH2, MSH6, PMS2, MUTYH or any other well described mismatch repair gene. There were three separate BRCA2 mutations found (BRCA2-p.1605fs*11, BRCA2-p.N1784fs*7, and BRCA2 R2888H). Germline testing was negative.


ASCO University
Re: GI Malignancies - dMMR (June 2019)
Jun 18, 2019 10:28 AM


Discussion Question

What are immunotherapy biomarkers for gastroesophageal cancers?


ASCO University
Re: GI Malignancies - dMMR (June 2019)
Jun 25, 2019 10:00 AM

Faculty response on behalf of Dr. Shi (Pathologist)

dMMR/MSI-H status and PD-L1 expression have been used as predictive biomarkers for checkpoint inhibitor therapy since the FDA approved pembrolizumab for the treatment of MSI-H or PD-L1-positive unresectable or metastatic gastric and esophageal cancers. While MSI status has been shown to be highly predictive of response to PD-1 blockade in several clinical trials, the value of PD-L1 expression in selecting patients for checkpoint inhibitors in esophageal and gastric cancer needs to be further investigated.

Approximately 40% of gastric/esophageal cancers express PD-L1. Unlike other malignancies (i.e., non-small cell lung cancer), PD-L1 expression in gastric/esophageal cancers is mainly observed in immune cells and not in tumor cells. The combined positive score (CPS), which takes into account the PD-L1 expression both in the tumor and tumor-associated immune cells, has been developed and refined for gastric and esophageal cancers. CPS is calculated by dividing the total number of PD-L1 positive cells (including tumor and immune cells) by the total number of viable tumor cells. A CPS ≥ 1 as determined by an FDA-approved companion diagnostic test (the Dako PD-L1 IHC 22C3 PharmDx Assay) is considered positive. A low overall response rate (ORR) to checkpoint inhibitors has been reported when using CPS cutoff of 1. Many studies are ongoing to explore if the ORR can be improved by using a higher CPS cutoff.

In addition, response to immune checkpoint inhibitors has been shown to correlate with high TMB. Therefore, TMB has been of interest as a potential predictive biomarker for immunotherapy in many malignancies. High TMB is frequently observed in dMMR/MSI-H tumors. It is also seen in microsatellite stable tumors, but infrequently. TMB has traditionally been assessed by whole-exome sequencing on both tumor and normal tissue. Recent studies have demonstrated that it can be accurately estimated by using large cancer gene-targeted sequencing panels. TMB is generally obtained by normalizing the total number of somatic nonsynonymous mutations to the total number of megabases sequenced. Currently there are no standard TMB measurement methodologies and thresholds; high TMB has been defined at various thresholds and with different NGS targeted panels. For example, Foundation Medicine Panel targets 315 genes (1.2 Mb), uses tumor tissue only, and has a cutoff of 10 mutations per Mb of DNA. MSK-IMPACT panel sequences 468 genes (1.5 Mb), uses both normal and tumor tissues, and has no universal threshold for all tumor types. Typically, at least 20% tumor cellularity and at least 50 ng of FFPE DNA is needed. 


ASCO University
Re: GI Malignancies - dMMR (June 2019)
Jun 25, 2019 10:02 AM

Faculty response on behalf of Dr. Cannon (Medical Oncology)

Dr. Shi did an excellent job of responding to this question. The predictive nature of MSI status is the most robust. Pembrolizumab is approved for refractory PD-L1 expressing gastric or GE junction adenocarcinoma with a combined positive (CPS) score (which Dr. Shi explained above) of 1 or higher. This was based on KEYNOTE 059, which noted an ORR of 1.3% in this population. Responses did occur in some MSS patients as well, but less frequently.

High expressers of PD-L1 may benefit from using immune checkpoint inhibitors as frontline therapy. KEYNOTE 062 was presented at this past ASCO by Tabernero et al. For patients with a PD-L1 CPS of 10 or more, the median OS for front-line Pembrolizumab was 17.4 months as opposed to 10.8 months for chemotherapy.

So, immunotherapy should be given, at some point, for appropriate patients with  MSI high and high PD-L1 expressing tumors. For the rare case of a PD-L1 nonexpressing MSS patient with a high mutational burden, I would still search for a clinical trial for an immune checkpoint inhibitor.

At our molecular tumor board, we also discuss some less well-established factors that may predict for responses to immune checkpoint inhibitors. For example, ATM mutations may be predictive in prostate cancer and the TAPUR trial is evaluating the use of combined immunotherapy in deleterious mutations in genes that have a role in homologous repair such as BRCA1, BRCA2, ATM, etc. Additionally, PRBM1 may be predictive as demonstrated in RCC.


ASCO University
Re: GI Malignancies - dMMR (June 2019)
Jun 25, 2019 10:09 AM

Patient Case 2 Update

Mr. Scott started nivolumab and ipilumumab in November 2017 as part of the Targeted Agent and Profiling Utilization Registry (TAPUR) clinical trial. After three cycles, he had a partial response.  He paused treatment to have back surgery for a severe discitis that had started prior to therapy, and then continued on trial. After six cycles, the next CT scan showed further improvement, with the 4b lesion shrinking to 7mm x 7 mm and the peripancreatic lymph node diminishing to less than 2 cm in largest diameter. He has been on nivolumab since early 2018 with no disease progression as of May 2019. Figure 2 shows the most recent view of the peripancreatic lymph node (measuring 12.7 mm in April 2019).