Noninvasive Molecular Detection of Colorectal Neoplasia

Transcription

1 Disclosures Noninvasive Molecular Detection of Colorectal Next Generation Approaches David Ahlquist August 7, 29 Relationship with Exact Sciences Mayo Clinic is equity investor Potential for future royalties to Mayo & Dr. Ahlquist for licensed IP Dr. Ahlquist serves as Senior Scientific Advisor 2 Effective Test Detection Effective Detection = S x C x A S = C = compliance A = access Screening Tools: Multiple Choices Accuracy Cost Invasive Access Stool blood Low Low No Unlim Flex sig Mod Mod Yes Lim Colon x-rayy Mod Mod Yes Lim Colonoscopy High High Yes Lim CTC High High Yes Lim Stool DNA M-H Mod No Unlim Other???? Ideal High Low No Unlim 3 CP Next Generation Molecular Tests Potential Advantages Noninvasive Avoidance of diet, medication, or cathartic preparation Off-site collection Wide accessibility Expanded value Improved effective detection of CRN Pan-GI cancer screening Improved Effective Detection of CR 5 1

2 Colorectal Cancer Incidence and Remaining Lifetime by Age (USA) Incidence e/1, Field changes Cancers Precancers Age (yr) Avg Rema aining Yrs 7 CP Correct Target Lesions Curable stage cancer Advanced adenoma Serrated polyps* Precursor for >2% of CRCs Elderly, R>L colon Sinister natural hx BRAF mut, methylation Often ignored as outcome Detection of precursor essential for CRC prevention ention *Sessile serrated adenoma images courtesy of Dr. Won Kee Song, Mayo Clinic 9 Broadly Informative Markers Well-established DNA alterations in CRC and precursor tissue Advanced adenoma APC 7-8%, KRAS 4-5%, meth genes >7% Serrated polyps BRAF ~7%, KRAS ~15-3%, meth genes >9% Several marker combinations detect ~1% of adv adenomas on tissue specimens, e.g APC, KRAS, meth vimentin (Ann Intern Med, 28;149:441)

3 Marker Stability in Stool Markers variably metabolized in stool DNA instability accounted for suboptimal neoplasm detection rates in multicenter with 1 st generation tests 14 Stabilization 12 1 buffer preserves 8 6 marker yield Median Percentage of Long DNA Left 2 Zou et al. CEBP 26; 15: Day Day 1 Day 3 Day 8 Days at Room Temperature 1 mm EDTA 16 mm EDTA High Analytical Sensitivity Stool DNA with adv adenomas Median level M:WT ~.5% Stool DNA assay detection limit: 1 st gen: >1% Next gen: <.1% Higher analytical sens yields higher clinical sens for adv adenomas Zou et al. Gastroenterology 29;136; Detection of CR Neoplasms by Next Generation Stool DNA Assay Zou et al, DDW 29 Subjects: 1 normals, 1 CR neoplasia Marker panel: Alu, KRAS, APC, mbmp3 Detection at 9% specificity Sensitivity,% Cancer Adv Ad >1cm 63 >2cm 78 >3cm

4 Full Anatomic Coverage of Colorectum Structural screening approaches biased toward left-sided CRC mortality reduction Flex sig (Selby et al. NEJM 1992;326:653) 653) Colonoscopy (Baxter et al. Ann Intern Med 29;15:1; Gupta et al. CGH 25; 3:15) Ideal test as complement or alternative to structural screening would detect both R and L 19 Stool Test Detection of CRN (CA + Adv Ad) in Screen Setting: Effect of Tumor Site % Prox Dist p-value Morikawa (n=21,85)* FIT <.1 1 Ahlquist (n=4,482)** Hemoccult HemoccultSensa DNA (SDT2) NS *Gastroenterology 25;125:422 (colonoscopy as gold std) **Ann Intern Med 28;149:441 (colonoscopy as gold std, unbuffered stools) 2 Colorectal Cancer Detection Gastroenterology 2;119:1219 Multi-target DNA test Positive ( ) 2 Negative ( ) 2 Sensitivity 2/22=91% CP Stool vs Other Media? Markers likely shed into colorectal lumen (stool) sooner than into circulation DNA have been detected in plasma and in urine at rates 6-8% with CRC but -2% with adv adenomas RNA expression assays in circulating mononuclear cells may detect adv adenomas (Arber et al. DDW 29) Other media (e.g. rectal mucin)? 23 Altered APC in Plasma % (+) % M:WT Normals (1) Large adenoma (11) 9.2 Cancer Dukes A (8) 63.4 Dukes B (8) Dukes D (6) Diehl et al, PNAS 25;12:

5 Screen Test Frequency Must accommodate natural hx Short-interval testing increases program e.g. 5% with 1 st screen, increases to 75% with 2 nd screen,. Long-interval testing may compromise program Effective Test Penetration Effective Detection = S x C x A S = C = compliance Test Penetration A = access High compliance requires userfriendly test features Wide access requires high throughput and broad distribution Pan-detection of GI by Stool DNA Testing Pan-Detection of GI Shared exfoliative biology Aggregate rather than single site prevalence Potential for tumor site prediction using multi-marker panel Taylor et al. DDW 29 Demonstrated feasibility 3 5

6 Pan-Detection of GI Neoplasms by Stool DNA Testing (Zou et al. DDW 29) Summary % Pos Oral Esoph Stom Panc Bil/GB CR Panc CR Cancer Precancer 31 Next generation molecular tests hold promise for expanded value Improved CRN detection Pan-GI cancer screening Technical features can be engineered to deliver optimal performance Validation and comparison studies needed 32 Other Future Approaches to Colorectal Cancer Screening Video Capsule Endoscopy Van Gossum et al. NEJM 29; 361:3 Self-intubating colon scopes (e.g. Aeroscope) Multicenter European study N = 328 referred patients Swallowed video capsules Colonoscopy on all VCE completed in 93% within 1 hrs Accuracy for advanced adenomas Sensitivity 73% Specificity 79%