to cancer research? Jason B. Fleming, MD Chair, GI Oncology H. Lee - PowerPoint PPT Presentation

How can a surgeon contribute to cancer research? Jason B. Fleming, MD Chair, GI Oncology H. Lee Moffitt Cancer Center

No Relevant Disclosures • None

Projected Increase in Deaths from Pancreatic Cancer in the US 2015 www.pancan.org

Surgery offers Hope of Cure

Contemporary Surgical Outcomes • National Cancer Data Base (~76% of cases in US) • 4739 patients received PD in 587 hospitals from 2010 and 2011 Institutions categorized into quartiles based on PD case volume • 30-day and 90-day mortality >6% Cases per Year: 1-5 6-13 14-25 >25 Kulu and Conrad, unpublished, 2016.

10,000 Hours and Pancreatic Surgery The Surgical “Learning Curve” EBL OR Time Hospital Stay MDACC surgeons (n=3) ~ 60 pancreas resections are necessary before a significant reduction in EBL, OR time and hospital stay is realized. Tseng, Surgery 141: 456 (2007).

Improvements with Resection of PDAC Progressively Limited Return on the “investment’ of Pancreas Surgery • 1970-2006 Johns Hopkins University • Postoperative complications: 38% • Mortality rate: 1% • Median OS: 19 mo; 5-yr OS: 20% in 2000’s Winter, J Gastrointest Surg 2006.

The Preoperative Therapy Approach Treatment Treatment Break Phase Best-evidence Therapy- but based upon incomplete information for individual patient Repeat OR Staging Staging Dropout “Could not make Establish Dx the leap across” Varadhachary et al., Ann Surg Onc 2006 Katz et al., JACS 2008

Observation: Patients fail mainly when systemic therapy ineffective. Preop Rx Follow Up Survivors 60-70% 35% 30-40% 65% Surgery Recurrence Dx Resistant Clinicopathologic Information Drop out

Observation: Distant Failure Predominates Ann Surg Oncol. 2009 April; 16(4): 836 – 847.

Observation: Ideal systemic therapy and local control can cure Gem/Cis Followed by XRT (50.4Gy) with Bevacizumab Erlotinib Capecitabine. Path CR 0/18 Nodes Pos. NED for >5yrs

Questions on Whipple Number 50…. What is something How can we get cool that I can do more research with all these going in pancreatic cancer pancreatic tumors I am taking out every week? cancer?

Dir irect Xenografts from Surgic ical Specimens Tumor Harvest Tumor Preparation Original Tumor Xenograft Tumor Tumor Implantation Subsequent Xenograft Original Tumor Xenograft Tumor Kim, et al. Nature Protocols, 2009.

Patient-Based Resource Building Preop Rx Follow Up Survivors 60-70% 35% 30-40% 65% Surgery Recurrence Dx Resistant Clinicopathologic Information 1 o Reagents blood sample = tumor sample Banked tumor sample 2 o Reagents Banked Banked TMA Tumorgraft Repository Post-treatment Immortalized Organoids Ex vivo RNA/DNA RNA/DNA PBMCs Cell Lines Testing Platform

Learning Strategy for Improvement Dx Preop Rx Sx Follow Up Novel Rx New Patients Biomarker Novel Rx Common Platform

PDAC PDX Program at MDACC PATX (Pancreatic tumor xenografts) Storage F3 DNA F2 RNA Xenografts F1 Protein TMA EVOC/LTSA Tumor tissue DNA Cell lines PATC (Pancreatic tumor EVOC: Ex Vivo Organotypic Culture LTSA: Live Tissue Sensitivity Assay cell line)

Summary of PDAC PDX Reagents (02/2008-05/2016) U n s u c ce s s fu l 7 3 /2 0 4 S u c c e s s fu l 1 3 1 /2 0 4 Number of Patients consented/implanted 220/204 cases Number of Xenograft tumors 131 (64%) T o ta l= 2 0 4 P rim a ry s ite (8 7 .7 % ) PDAC cell lines from xenograft tumors 23 L iv e r m e t (6 .8 % ) B o n e m e t (1 .5 % ) Characterized cell lines* 14/23 M a lig n a n t a s cite (0 .7 % ) P e rito n e a l m e t (1 .5 % ) L u n g m e t (0 .7 % ) Fingerprinted 14/23 L ym p h n o d e m e t (0 .7 % ) Collaborating PIs 60 PIs or Labs P rim a ry site (8 7 .7 % ) Annual new xenografts 15 to 25 cases T o ta l= 1 3 1 P D X s

Observation: Time is your enemy What Happens in Operating Room/Pathology: 1. Blood supply ligated 2. Specimen removed 3. Sits on back table 4. Circulator nurse 5. Path Tech paged 6. Path Tech arrives 7. Delivered to Path Several 8. Sits on table Hours! 9. Specimen registered 10. Sits on Dissection table 100 Halling, et al. 2003 % TUNEL pos Nuclei/HPF 11. Tumor dissected 80 12. Research samples obtained 60 13. Research Tech paged 40 14. Research Tech arrives 15. Research Tech takes specimen 20 16. Downstream experiment 0 0 30 60 120 180 Goal: engraftment time <30 minutes Time from Specimen Removal (minutes) Surgical Advantage

Observation: Tumorgraft growth reflects tumor biology Tumorgraft growth Yes No Median survival 613 2067 (days) • • Disease-free survival Overall survival (n=14) (n=14) (n=56) (n=56) Ann Surg Oncol (2015) 22: 1884-1892

Observation: Plastic Changes Cancer Cell Populations Patient66 Patient43 Patient50 Patient53 PATX43 PATX50 PATX53 PATX66 MDA-PATC43 MDA-PATC53 MDA-PATC66 MDA-PATC50 Tumorgraft E-Cadherin N-Cadherin Primary PC Cell Lines 100% 100% 50% 50% 0% SUB-… SUB-… SUB-… SUB-… 0% SUB-… SUB-… SUB-… SUB-… PT43 PATX43 PT50 PATX50 PT53 PATX53 PT66 PATX66 PT43 PATX43 PT50 PATX50 PT53 PATX53 PT66 PATX66 Kang, et al. Lab Investigation. 2015.

Technical Advance: Cryostorage and Reanimation of PDX Ivanics, et al. 2017.

Technical Advance: Fine Needle Aspirates Generate PDX Xenograft Prior Engraftment rate with Engraftment rate source therapy traditional method with biopsy method Primary Yes 3/5 0/5 Primary No 3/5 1/5 Met Yes 5/5 5/5 Primary No 1/5 0/5 Primary Yes 0/5 0/5 Primary Yes 0/5 0/5 Primary Yes 4/5 0/5 Primary No 4/5 3/5 Met Yes 3/3 N/A Primary Yes 0/5 0/5 � Roife.Surgery. 2017.

Questions on Whipple Number 150…. I need some people How can we use way smarter than the xenografts to me…. learn more?

Subtyping by Gene Transcription Signature Material: RNA extracted from microdissected PDAC cells + cell lines Classical QM-PDA Exocrine 62 gene signature Classical is K-ras-driven Survival of 27 Resected Cases (UCSF) Hepatogastroenterology 55 , 2016 – 2027 (2008) Nat Med. 2011 Apr;17(4):500-3.

PDX Transcription Profiles Collisson Moffitt Classical PDX provide excellent quality reagents for reproducible WES QM-Basal Chris Bristow and Tim Heffernan PDX

Inference of regulators of basal /classical tumors • Leverage compendium of PDX model data to infer pathway activity • virtual inference of protein activity by enriched regulon analysis (VIPER) VIPER TCGA PDAC-specific RNAseq profilesA data interactome Subtype VIPER classification analysis Subtype associated regulators Alvarez et al. Nat Gen , 2016 Classical regulators Basal regulators

Examples of context-specific regulators (VIPER) PDX inferred GATA6 suppresses EMT activity Basal Classical KLF5 regulates epithelial genes Diaferia et al , EMBO, 2016 Martinelli et al , Gut, 2015

Goal: Cross-cutting Searchable Data Platforms Imaging/Biophysical Clinicopathologic Tumor Sensitivity Molecular

Questions on Whipple Number 250…. Are there any I need even more clinically available people way smarter data that could than me…. help?

Imaging Subtypes Observed Patient 1 Patient 2 Koay, Truty, Cristini, et al. J Clin Invest. 2014 Apr 1;124(4):1525-36.

Low and High Delta Subtypes A Overall survival (OS) stratified by delta measurement for patients HU B Segmentation for delta Example: Histogram High delta Low delta Koay, unpublished, 2017.

Summary Stromal content/tumor cell proliferation ≈ Stability Parameter ( L) Koay, unpublished, 2017.

Imaging Characteristics Are Reflected in the Growth Kinects of Tumorgrafts High Delta r=0.4752 R 2 =0.2258 P=0.0080 Low Delta r=0.1288 R 2 =0.01659 P=0.7832

PATX-69 PATX-50 PATX-118 PATX-66 PATX-102

Texture analysis of MR images of PATX tumors IBEX (open infrastructure software platform, imaging biomarker explorer) Hi D Lo D • Survival curve of 6 xenografts

High Low Mutation Classification and delta delta Imaging Subtype Proportion with ( B ) Koay, unpublished, 2017.

Example: PATX-69 Tissue segmentation - Red: tumor - Green: stroma - Blue: normal pancreas

Observations: Stability of stroma with PDX • Representative images INDIVIDUAL PATIENT TUMOR • Comparison of stroma content in tumor from patients and PHENOTYPE DETERMINES from patient-derived xenografts DEGREE OF FIBROSIS 100 Collagen Area Fraction % 80 60 PATX11 40 PATX7 20 PATX4 PATX1 0 F0 F1 F2 F3 F0 F1 F2 F3 F0 F1 F2 F3 F0 F1 F2 F3 Tumor Generation

Low and High Delta Subtypes and Stromal Collagen PDX PATIENT CASES Low delta High delta Low delta High delta Tissue category CT scan H&E pathologist) Stroma score (by High delta Low delta Imaging Phenyotype Koay, unpublished, 2017

Nuclei of cancer cells from high delta tumors are more elongated suggesting aggressive biology • • Test set (12 cases) Validation set (17 cases)

Stability of nuclei morphology of cancer cells with passaging F1 F2 F3 F4 F5 • Representative H&E stained images (PATX 118) 1 2 3 5 4 4 1 2 3 5 3 1 2 3 4 1 2 4 2 3 5 1 2 4 5 3 1 2 1 3 5

PDAC Cell Shape and Imaging Phenotype 1.0 Koay, unpublished, 2017.