CON: C Can ctDNA as as a a Mar arker o of f Minimal Residua - PowerPoint PPT Presentation

AL B BENSON III, MD, FACP, FASCO PROFESSOR OF MEDICINE ASSOCIATE DIRECTOR FOR COOPERATIVE GROUP ROBERT H LURIE COMPREHENSIVE CANCER CENTER OF NORTHWESTERN MEDICINE CON: C Can ctDNA as as a a Mar arker o of f Minimal Residua dual D Disea ease e Be Used t d to Direc ect A Adjuv uvant Therapy i in Co Colon Can Cancer?

MD Anderson 2 Relative potential of ctDNA applications for cancer patients Initial molecular Serial molecular profiling profiling Other Minimal residual disease Treatment monitoring

Incidence of ctDNA based on tumor histology

NGS Assay Assay with 197 genes; at least one mutation detected 99.3% of tumor tissue 57% sensitivity for recurrence; 100% specificity • Stage III (16% prevalence of ctDNA+) • Stage II (5% prevalence of ctDNA+) HR 54.4 95% CI: 9.5-311.7 p<0.0001 Ho How d do w we i e improve o outcomes f for HR 20.0 95% CI: 5.9-67.8 thes ese p patients ts? p<0.0001 Diehn et al ASCO ‘17

NGS Assay Assay with 197 genes; at least one mutation detected 99.3% of tumor tissue 57% sensitivity for recurrence; 100% specificity • Stage III (16% prevalence of ctDNA+) • Stage II (5% prevalence of ctDNA+) Do w we n e need eed a as m much t therapy f for thes ese p patients ts? HR 54.4 95% CI: 9.5-311.7 p<0.0001 HR 20.0 95% CI: 5.9-67.8 p<0.0001 Diehn et al ASCO ‘17

NCI / Colon Cancer Task Force ctDNA Workshop • Met at ASCO ’18, with goal of evaluating ctDNA utility, making recommendations to industry, and identifying NCI opportunities in 4 key areas: • Use of ctDNA for management of minimal residual disease, Monitoring of metastatic disease, Acquired resistance against targeted therapies, and Specific issues relevant to rectal cancer. 6

MD Anderson Recommendation: Assay characteristics necessary for routine testing for minimal residual disease • For escalation applications • High specificity/PPV, even at the expense of lower sensitivity • PPV should be >90-95%; No more than 1 in 10 false positives • Turnaround time will be critical to make real-time decisions. May require non- personalized approaches. • Having matched tumor available – which would require enrollment at the time of surgery • Multi-gene would be preferred to allow broad capturing of potential patients. NCI / Colon Cancer Task Force Workshop ‘18

Use of ctDNA to guide escalation of adjuvant therapy • Need additional data: clearance of ctDNA with chemotherapy • Is the presence of ctDNA after surgery purely prognostic? • Is there an opportunity to change outcomes with therapy? • Can standard chemotherapy clear these patients? • FOLFOX in stage III disease • Novel therapies may be needed Tie, et al. ASCO 201 NCI / Colon Cancer Task Force Workshop ‘18

Phase 2 Adjuvant Studies: cfDNA to facilitate drug development in early stage tumors Cu Current L Landsc scape Potential A App pproach ch • No mechanism to obtain proof- • Validate cfDNA as a surrogate of-concept for adjuvant efficacy marker of relapse • Requires commitment for • Perform proof-of-concept n>1000 patient studies through small phase 2 studies looking at cfDNA clearance • Result: Very few adjuvant studies • Result: Reduced risk and conducted, and only with increased innovation for adjuvant approved therapies in mCRC studies

ARGUMENTS AGAINST CTDNA

Concordance between tumor and liquid biopsies for mutational analysis

Commercially available testing kits

Courtesy of Dr. Julie Lang, USC

Driver mutations seen in non-cancer patients

Sensitivity of testing remains low and unreliable

Not all tumors shed ctDNA at same rate

Concordance with tissue NGS remains poor • Concordance of genomic alterations between two commercially available ctDNA (guardant) and tissue biopsies (tempus) was compared in 45 patients with breast cancer using paired next-generation sequencing tissue and ctDNA biopsies. • Across all genes, concordance between the two platforms was 91.0% to 94.2%. • genomic alterations in either assay (e.g., excluding wild type/wild type genes), concordance was 10.8% to 15.1% with full plus partial concordance of 13.8% to 19.3%. • Concordant mutations were associated with significantly higher variant allele frequency. • Over half of mutations detected in either technique were not detected using the other biopsy technique Chae et al, 2017. Mol Can Thera

MD Anderson 19 Conclusions • Disease monitoring applications now require prospective studies • Early detection of resistance does not necessarily result in better outcomes • Approval of this indication will require well-designed studies in discrete indications • Minimal residual disease applications have tremendous opportunity • Requires larger, prospective cohorts • Great opportunities for novel drug development • Private/public partnerships will likely be required

NRG Stage II Adjuvant Study: CR1643 Evaluating early intervention for Minimal Residual Dz Van Morris Primary objective: Clearance of cfDNA (to undetectable levels) for patients cfDNA+ at randomization Morris, Kopetz

Slide 27 Presented By Ryan Corcoran at 2019 ASCO Annual Meeting

Thanks to Scott Kopetz and Aparna Kalyan for use of their slides

EXTRA SLIDES

• Adjuvant t t therap apy c can c clear ctDNA • ctDNA cleared in 50% (9 of 18) patients who completed the entirety of adjuvant chemotherapy with a 5-fluorouracil/oxaliplatin combination. • Persistence of ctDNA after chemotherapy was associated with a worsened recurrence risk (HR 7.1, p < .001). • Se Seria rial m l monitorin ing w will i increase s sensitiv ivit ity • Nearly 10% of the initially ctDNA-negative population converted to a positive status after chemotherapy 24 Tie et al , ASCO 2018;36(15_suppl):351

“Phase II” Adjuvant Studies Phase I e II • Endpoint of clearance of ctDNA, where this is ne necessary but no y but not s t suf ufficient for 1 ° endpoint: ctDNA+ population Novel cure DFS (~100% rate of Interventio • High event rate, so feasible for DFS -or- radiographic Clearance of ctDNA recurrence) n endpoint as well Phase III III • Holy Grail: Can we use ctDNA as an FDA- “De-risk” a Novel intervention 1 ° endpoint: traditional R approved surrogate endpoint for DFS or OS phase III SOC registration of novel therapies? Integral Novel intervention 1 ° endpoint: biomarker: R DFS or OS ctDNA+ SOC 25 population

Use of ctDNA to guide de-escalation of adjuvant therapy • The group felt that sensitivity of 90-95% was required to guide de-escalation (vs. escalation) • Need to define how disease characteristics may influence sensitivity of the test • eg. peritoneal disease may not shed ctDNA at same rate • These studies are typically larger due to a non-inferiority design • International consortia are needed NCI / Colon Cancer Task Force Workshop ‘18

Concordance with tissue NGS remains poor Concordance of genomic alterations in tissue and ctDNA categorized by potential functionality. Shown here is tissue (T) and ctDNA. T + ctDNA+, concordant in both platforms. T − ctDNA+: variant found in ctDNA, but not tissue. T + ctDNA −: variant found in tissue, but not ctDNA.. Chae et al 2017, Mol Can Therap