American Surgical Association Annual Meeting
Search Meeting Site Only
 
Main ASA Website
Current Meeting Home
Final Program
Past & Future Meetings

 

 

Back to 2016 Annual Meeting


A Highly Predictive Model for Diagnosis of Colorectal Neoplasms Using Plasma MicroRNA: Improving Specificity and Sensitivity
Jane V Carter*1, Jonathan Rice*1, Henry Roberts*1, Shesh N Rai*1, Ziad Kanaan*2, Susan Galandiuk1
1University of Louisville, Louisville, KY;2Wayne State University, Detroit, MI

OBJECTIVE(S):
Develop a plasma-based microRNA(miRNA) diagnostic assay specific for colorectal neoplasms, building upon prior work(ASA-2012).
Colorectal neoplasms (colorectal cancer[CRC] and advanced adenoma[CAA]) frequently develop in individuals at ages when other common cancers also occur. Current screening methods lack sensitivity, specificity, and have poor patient compliance.
METHODS:
Plasma was screened for 380miRNAs using microfluidic array technology from a “training” cohort of 60 patients, (10 each) control, CRC, CAA, breast(BC), pancreatic(PC) and lung(LC) cancer(Table). We identified uniquely dysregulated miRNAs specific for colorectal neoplasia (p<0.05, false discovery rate:5%, adjusted α=0.0038). These miRNA were evaluated using single assays in a “test” cohort of 120 patients. A mathematical model was then developed to predict sample identity in a 150 patient blinded “validation” cohort using repeat-sub-sampling validation of the testing dataset with 1000 iterations each to assess model detection accuracy.
RESULTS:
Seven miRNAs (miR-21, miR-29c, miR-122, miR-192, miR-346, miR-372, miR-374) were selected based upon p-value, AUC, fold-change, and biological plausibility. AUC for test cohort comparisons were 0.91, 0.79 and 0.98, respectively(Table). Our mathematical model predicted blinded sample identity with 69-77% accuracy in comparison #1, 66-74% accuracy in comparison #2, and 86-90% accuracy in comparison #3(Table).
CONCLUSIONS:
Our plasma miRNA assay and prediction model differentiates colorectal neoplasia from patients with other neoplasms and from controls with high sensitivity and specificity compared to current clinical standards. This appears promising for colorectal neoplasia identification.
Study Design and Accuracy of Diagnostic Model
Study DesignTest Cohort Comparison (n=120) Area Under the Curve (AUC) (95% CI) for Selected miRNAs in Test CohortValidation Cohort Comparison (n=150)Accuracy of Model to Predict Blinded Sample Identity in Validation Cohort
Training Cohort
380 miRNA assessed using microfluidic arrays
n=60
10 each
Control, CRC, CAA, BC, PC, LC
#1: Any Neoplasia vs Control
n=100 vs 20
0.91 (0.85-0.96)#1: Any Neoplasia vs Control
n=125 vs 25
69-77%
Test Cohort
7 miRNA dysregulated in CR neoplasia assessed using single miRNA assays
n=120
20 each
Control, CRC, CAA, BC, PC, LC
#2: CRC Neoplasia vs Other Cancers
n=40 vs 60
0.79 (0.70-0.88)#2: CRC Neoplasia vs Other Cancers
n=50 vs 75
66-74%
Validation Cohort
7 miRNA dysregulated in CR neoplasia assessed using single miRNA assays
n=150
25 each
Control CRC, CAA, BC, PC, LC
#3: CRC vs CAA
n=20 vs 20
0.98 (0.96-1.0)#3: CRC vs CAA
n=25 vs 25
86-90%


Back to 2016 Annual Meeting


© 2022 American Surgical Association. All Rights Reserved. Privacy Policy.