Article Text

Download PDFPDF

Detection of microsatellite instability-high (MSI-H) by liquid biopsy predicts robust and durable response to immunotherapy in patients with pancreatic cancer
  1. Sakti Chakrabarti1,
  2. Leslie Bucheit2,
  3. Jason Scott Starr3,
  4. Racquel Innis-Shelton4,
  5. Ardaman Shergill5,
  6. Hiba Dada2,
  7. Regina Resta6,
  8. Stephanie Wagner7,
  9. Naomi Fei8 and
  10. Pashtoon Murtaza Kasi9
  1. 1Hematology-Oncology, Medical College of Wisconsin, Milwaukee, Wisconsin, USA
  2. 2Guardant Health Inc, Redwood City, California, USA
  3. 3Department of Hematology/Oncology, Mayo Clinic Florida, Jacksonville, Florida, USA
  4. 4University of Alabama, Birmingham, Alabama, USA
  5. 5University of Chicago Pritzker School of Medicine, Chicago, Illinois, USA
  6. 6New York Oncology Hematology PC, Albany, New York, USA
  7. 7Columbus Regional Health Care Center, Columbus, Indiana, USA
  8. 8The University of Iowa Healthcare, Iowa City, Iowa, USA
  9. 9Weill Cornell Medicine, New York, New York, USA
  1. Correspondence to Dr Sakti Chakrabarti; SChakrabarti{at}


Clinical trials reporting the robust antitumor activity of immune checkpoint inhibitors (ICIs) in microsatellite instability-high (MSI-H) solid tumors have used tissue-based testing to determine the MSI-H status. This study assessed if MSI-H detected by a plasma-based circulating tumor DNA liquid biopsy test predicts robust response to ICI in patients with pancreatic ductal adenocarcinoma (PDAC). Retrospective analysis of patients with PDAC and MSI-H identified on Guardant360 from October 2018 to April 2021 was performed; clinical outcomes were submitted by treating providers. From 52 patients with PDAC +MSI-H, outcomes were available for 10 (19%) with a median age of 68 years (range: 56–82 years); the majority were male (80%) and had metastatic disease (80%). Nine of 10 patients were treated with ICI. Eight out of nine patients received single-agent pembrolizumab (8/9), while one received ipilimumab plus nivolumab. The overall response rate by Response Evaluation Criteria in Solid Tumors was 77% (7/9). The median progression-free survival and overall survival were not reached in this cohort. The median duration of treatment with ICI was 8 months (range: 1–24), and six out of seven responders continued to show response at the time of data cut-off after a median follow-up of 21 months (range: 11–33). Tissue-based MSI results were concordant with plasma-based G360 results in five of six patients (83%) who had tissue-based test results available, with G360 identifying one more patient with MSI-H than tissue testing. These results suggest that detecting MSI-H by a well-validated liquid biopsy test could predict a robust response to ICI in patients with PDAC. The use of liquid biopsy may expand the identification of PDAC patients with MSI-H tumors and enable treatment with ICI resulting in improved outcomes.

  • Immunotherapy
  • Gastrointestinal Neoplasms
  • Programmed Cell Death 1 Receptor
  • Tumor Biomarkers

Data availability statement

No data are available. Not applicable.

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Pancreatic ductal adenocarcinoma (PDAC) is an aggressive malignancy with a 5-year overall survival rate of less than 10%.1 In the USA, approximately 60,000 patients are diagnosed with PDAC annually, most presenting with metastatic disease.2 While there has been a rapid increase in the development of targeted therapies in other cancer types, PDAC generally lacks a targetable alteration.2 Recent literature reported remarkable antitumor activity of immune checkpoint inhibitor (ICI), irrespective of tumor type and checkpoint inhibitor used, in patients harboring tumors with a high level of microsatellite instability (MSI-H) as a result of deficient mismatch repair (dMMR).3 The detection of MSI-H status in patients with pancreatic tumors may provide a unique opportunity for treatment with ICI, although the prevalence of MSI-H signature in patients with PDAC is quite low (<1%).4

MSI-H status can be assessed in cancer patients by genomic profiling and has historically been performed on tissue specimens. However, several well-known barriers to genomic profiling of tumor tissue exist, including tissue insufficiency or the inability to perform a tissue biopsy. Additionally, the invasive procedure involved in obtaining a tissue biopsy often poses challenges and may add to patient morbidity. A high degree of concordance between circulating tumor DNA (ctDNA)-based tumor genomic profiling and tissue-based tumor genomic profiling has been reported.5 As a result, the use of liquid biopsy for genomic profiling is rapidly gaining popularity6 and may be used to assess tumor genomic profiles in a low risk, timely fashion.

The use of well-validated liquid biopsies can not only help with the completion of genotyping, it can also rapidly identify MSI-H,6 which may offer expanded treatment opportunities in patients with a diverse group of tumor types, including pancreatic cancer, given tumor agnostic approvals. Previous studies reported significant antitumor activity of ICI in MSI-H gastric7 and patients with prostate cancer8 whose MSI-H status was detected by liquid biopsy testing. Here, we investigated the prevalence of MSI-H/dMMR tumors in a large cohort of patients with PDAC using a well-validated liquid biopsy assay and assessed if the dMMR/MSI-H signature detected by a plasma-based testing predicts robust and durable response to ICI.


Patients and samples

The Guardant central database was searched for patients with PDAC who had MSI-H tumors detected by a plasma-based liquid biopsy that assesses ctDNA, Guardant360 (G360), performed as a part of routine clinical care between October 1, 2018 and April 15, 2021. Clinicians providing care to the patients with MSI-H PDAC detected by G360 were contacted to obtain clinical data that included patient and tumor characteristics, treatment details, and outcomes. The data cut-off date was September 1, 2021.

Sequencing and analysis

G360 (Guardant Health, Redwood City, California, USA) is a commercially available 74-gene panel plasma-based tumor genomic profiling assay validated to detect a variety of genomic alterations, including MSI-H signature,7 single-nucleotide variants, indels, copy number alterations (amplifications and fusions)9 in cell-free DNA (cfDNA) from plasma of patients with solid tumors, including PDAC. G360 determines MSI-H status by sequencing 90 pan-cancer informative microsatellite loci in cfDNA and reports MSI-H status based on the number of unstable sites relative to a predetermined cut-off, as previously described.7 Reporting of MSI-H status was included in the G360 test result beginning September 27, 2018.


During the study period, over 6000 patients with PDAC had G360 as part of clinical care, 52 of whom had MSI-H identified for a prevalence of 0.8%. Clinical outcome data available for 10 of 52 (19%) patients were included in the final analysis. The cohort had a median age of 68 years (range: 56–82); 80% were male, and 80% of patients had metastatic disease (table 1).

Table 1

Characteristics of patients with MSI-H pancreatic cancer

The diagnosis of MSI-H was made by G360 in 7 of 10 (70%) patients. Most patients had KRAS or GNAS alterations, while more than half also had alterations related to homologous recombination repair identified by G360 testing in addition to the identification of MSI-H (table 2). Tissue analysis of MSI-H status by IHC was performed in 6 of 10 (60%) cases. All but one result was concordant; in the discordant case, immunohistochemistry (IHC) testing on the pancreatic body mass biopsy tissue failed to identify dMMR status, but G360 obtained 2 weeks before starting immunotherapy identified MSI-H. This patient received neoadjuvant therapy with an ICI combination (ipilimumab plus nivolumab) followed by surgery; the resected specimen confirmed complete pathological response.

Table 2

Co-occurring deleterious alterations and concordance of MSI-H identification between tissue immunohistochemistry and Guardant360 in patients with pancreatic ductal adenocarcinoma

ICI was administered in 9 of 10 patients, 7 of whom received ICI following the identification of MSI-H status by G360. The only patient who did not receive ICI passed away before being able to receive ICI, therefore, outcome analysis included 9 patients. Nearly all patients who received ICI received single-agent pembrolizumab (8/9), while one received ipilimumab plus nivolumab; however, patients received ICI across various lines of therapy . At the time of data cut-off, the overall response rate (ORR) by the Response Evaluation Criteria in Solid Tumors (RECIST) criteria was 77% (7/9) and 6 of 7 responders continued to show response at the time of data cut-off after a median follow-up of 21 months (range: 11–33) (figure 1). The median progression-free survival and overall survival were not reached in this cohort, where the median duration of ICI therapy was 8 months (range: 1–24). Among the nine patients treated with ICI, 7/9 (78%) were alive at the time of data cut-off.

Figure 1

Progression-free survival of nine evaluable patients with microsatellite instability-high pancreatic cancer treated with immune checkpoint inhibitor.


The present retrospective cohort study investigated if the detection of MSI-H with a liquid biopsy test (G360) in patients with PDAC predicts a robust response to ICI. In this study, PDAC patients who had MSI-H tumors detected by G360 showed a robust response to ICI, evidenced by an ORR of 77%. The responses were durable, with six out of seven responders experiencing disease control for a prolonged period. Additionally, the study demonstrated a high degree of concordance (83%) between the plasma-based and the tissue-based detection of MSI-H, with liquid biopsy able to identify MSI-H not identified on IHC testing in one patient. This is the first study to our knowledge reporting a robust response to ICI in patients with MSI-H PDAC in which MSI-H was detected by liquid biopsy.

MSI-H has emerged as a tumor-agnostic predictive biomarker for response to ICI, supported by several large prospective studies.3 10–12 The high antitumor activity of ICI in patients harboring MSI-H tumors irrespective of tissue of origin led to accelerated tissue-agnostic approval of pembrolizumab, an ICI that acts by blocking programmed death 1 (PD-1) receptor on the lymphocytes, by the Food and Drug Administration (FDA) of the USA for adult and pediatric patients with unresectable or metastatic MSI-H/dMMR solid tumors resistant to standard therapies. Subsequently, FDA approved pembrolizumab in treatment naïve patients with MSI-H metastatic colorectal cancer based on the KEYNOTE-177 trial data.10 However, a large data set confirming ICI activity in dMMR/MSI-H PDAC patients are unavailable. Although the immunosuppressive tumor microenvironment of pancreatic cancer characterized by a lack of T cells, an abundance of immune-suppressive myeloid cells, and dense desmoplasia is a cause for concern,13 preliminary data suggest that ICI has significant activity in dMMR/MSI-H PDAC patients. The pivotal study by Le et al investigating the antitumor activity of pembrolizumab in patients (n=86) with 12 different types of MSI-H/dMMR solid tumors had 8 patients of PDAC.11 In this study, the reported ORR was 62% (5/8) in patients with PDAC and 53% in the whole group.11 The remarkable activity of ICIs in MSI-H/dMMR pancreatic tumors has been reported in several case reports.14 15 Furthermore, the successful utilization of ICI as neoadjuvant therapy has been reported in several small studies and a case report in patients with localized or locally advanced dMMR/MSI-H tumors.14 16 17 Consequently, detecting MSI-H/dMMR in patients with PDAC can enable treatment with IO that often results in an improved outcome. The high response rate observed in this study is consistent with the previously reported studies with pancreatic and non-pancreatic tumors, supporting the feasibility of MSI-H status determination with plasma-based testing. It is important to mention that phase II KEYNOTE-158 study in which patients with chemotherapy-refractory MSI-H/dMMR advanced noncolorectal cancer received pembrolizumab reported a rather low response rate of 18%,18 a result discordant with most studies. The underlying cause of this discordance is unknown, although one might speculate if the prior therapies influenced the response rate.

It is unclear why some patients with MSI-H tumors do not respond to ICI. Two patients in the current cohort did not respond to ICI. Patient 1 had MSI-H confirmed on both G360 and tissue-based testing. However, this patient was tested for MSI late at progression and had worsening clinical status, ultimately leading to death after an early round of ICI. Patient 4 had MSI-H confirmed on tissue as well as G360 and was initially treated with single-agent pembrolizumab for 6 months with a mixed response but overall progressive disease by RECIST. At the time of data cut-off, the patient had received one dose of ipilimumab plus nivolumab, and early markers continued to be concerning for progression. Of note, his ctDNA was not cleared on subsequent G360 tests, showing active disease evolution even when actively treated with ICI. Concurrent alterations on G360 were identified in KRAS, ARID1A, MLH1, and TP53 (table 2); tissue NGS testing identified the same ARID1A, KRAS, and MLH1 alterations. Potential explanations for this patient’s non-response may include tumor heterogeneity, or an altered microenvironment, among others.

The prevalence of MSI-H/dMMR tumors in patients with PDAC appears to be low, around 1%,4 as seen in this study and others. Hu et al evaluated the mismatch repair status in 833 patients of PDAC using a next-generation sequencing assay where MSI-H tumor was detected in only 0.8% of patients.19 Another study analyzed 445 tumor specimens from patients with PDAC with an IHC-based assay and reported the presence of dMMR tumor in 1.6% of cases.20 Conversely, a single institution study reported dMMR tumors in 22% (24/109) of pancreatic tumor biopsies.21 The range of reported prevalence of MSI-H/dMMR tumors in PDAC patients is likely related to patient selection criteria in different studies. Overall, it appears that the prevalence of MSI-H/dMMR signature is low in patients with PDAC. Although the possibility of finding MSI-H tumors in PDAC patients is low, as it was in this study cohort, it is reasonable to test for MSI-H/dMMR status in all patients with PDAC as the identification of the MSI-H/dMMR status provides a unique opportunity for treatment with ICI that often leads to a robust response.

Insufficient tumor tissue in the biopsy for genomic profiling is a well-recognized barrier to genomic profiling and is frequently encountered in localized PDAC in which tissue collected by endoscopic ultrasound-guided fine-needle biopsy and aspiration yields inadequate samples in as high as 29% of cases.22 A recent trial showed that liquid biopsy might be able to overcome tissue-based genomic profiling challenges in advanced GI cancers, including PDAC, due to higher rates of successful genomic profiling, faster sample acquisition, and quicker result availability.22 Furthermore, IHC-based tests occasionally misclassify dMMR/MSI-H status23 24 as 5%–11% of MSI-H tumors may demonstrate intact MMR protein expression, likely related to retained antigenicity in otherwise nonfunctional MMR proteins.23 A well-validated liquid biopsy test can effectively fill these critical gaps as observed in one patient described in this study where an IHC-based test reported proficient MMR, but G360 identified MSI-H. This patient received neoadjuvant therapy with an ICI combination (ipilimumab plus nivolumab) followed by surgery, and the resected specimen confirmed complete pathological response. The intratumoral heterogeneity25 or retained antigenicity of the nonfunctional MMR proteins as described above23 may have contributed to the observed discordance in this patient. Such discordance between IHC and G360 has been described in a previous study in which G360 accurately identified MSI-H with IHC providing an incorrect result,6 highlighting the importance of using multiple methodologies to maximize the identification of patients with MSI-H tumors. Validation studies of G360 to determine the MSI-H status reported an overall accuracy of 98.4% and a positive predictive value of 95%,7 as well as multiple cohorts showing robust response to ICI, supports the utility of G360 to be used concurrently with comprehensive genomic profiling on tissue specimens and/or where tissue is insufficient or inaccessible. Communication with the coauthors revealed that payers agreed to pay for the prescribed ICI in patients who had MSI-H detected by liquid biopsy alone.

This study has several limitations, including the retrospective nature of this analysis, a small sample size, and the non-availability of treatment outcome information in a significant number of patients who had MSI-H tumors detected by G360. However, the robust and durable responses to ICI observed in this study among the patients with MSI-H PDAC provide confidence that prescribing ICI guided by plasma-based identification of MSI-H signature is appropriate.


In this small cohort of patients with PDAC, the detection of MSI-H by ctDNA testing was highly concordant to tissue-based testing and correlated with robust and durable responses to ICI. The use of a well-validated liquid biopsy assay may expand the identification of MSI-H tumors in patients with PDAC and enable treatment with ICI resulting in improved outcomes.

Data availability statement

No data are available. Not applicable.

Ethics statements

Patient consent for publication

Ethics approval

This retrospective study was approved by an institutional review board with a waiver of consent for the analysis of deidentified data.



  • Twitter @doctorC369, @drjasonstarr, @pashtoonkasi

  • Contributors Concept: SC, LB, PMK. Data collection: SC, JSS, RI-S, AS, RR, SW, PMK. Data analysis and interpretation: all authors. Visualization: HD. Manuscript writing: SC, LB. Manuscript editing and final review: all authors.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests SC: Advisory role- QED Therapeutics. Research support-Natera. Speakers Bureau- Natera. LB :Employee and shareholder of Guardant Health, Inc. HD: Employee and shareholder of Guardant Health, Inc.PMK: Consulting or Advisory Role- Taiho Pharmaceutical (Inst), Ipsen (Inst), Natera, Foundation Medicine, AstraZeneca, MSD Oncology, Tempus, Bayer, Lilly, Delcath Systems, Axiom Healthcare Strategies, IPBA, QED Therapeutics, Boston Healthcare Associates. Research Funding- Bristol Myers Squibb (Inst), Advanced Accelerator Applications (Inst), Array BioPharma (Inst), Celgene (Inst), Tersera (Inst), Boston Scientific (Inst).Travel, Accommodations, Expenses- AstraZeneca.

  • Provenance and peer review Not commissioned; externally peer reviewed.