SA201 - Diverticulitis Polygenic Risk Scoring in a Diverse Patient Population

Saturday, June 1, 2024

2:19 PM – 2:25 PM ET, USA

Location: Room 314-317

Abstract Podium Presenter(s)

Christopher J. Neylan, MD

General Surgery Resident
Hospital of the University of Pennsylvania
Hospital of the University of Pennsylvania
Philadelphia, PA, United States

All relevant financial disclosures have been managed and will be listed below.
This individual has no financial relationships with ineligible companies.

Author(s)

SA

Sarah Abramowitz, n/a

Medical Student
Zucker School of Medicine at Hofstra/Northwell, University of Pennsylvania Perelman School of Medicine
Philadelphia, Pennsylvania, United States

Disclosure information not submitted.
ML

Michael Levin, MD

Instructor of Medicine
University of Pennsylvania Perelman School of Medicine
Philadelphia, Pennsylvania, United States

Disclosure information not submitted.
JD

John DePaolo, MD

Resident, General Surgery
Hospital of the University of Pennsylvania
Philadelphia, Pennsylvania, United States

Disclosure information not submitted.
RJ

Renae Judy, MS

Bioinformatician
University of Pennsylvania
Philadelphia, Pennsylvania, United States

Disclosure information not submitted.
Jeffrey L. Roberson, MD

Resident
The Hospital of the University of Pennsylvania
The Hospital of the University of Pennsylvania
Philadelphia, PA, United States

Disclosure information not submitted.
SD

Scott Damrauer, MD

Associate Professor of Surgery and of Genetics
Perelman School of Medicine
Philadelphia, Pennsylvania, United States

Disclosure information not submitted.

Senior Author(s)

LM

Lillias Maguire, MD FACS

Assistant Professor of Surgery
University of Pennsylvania
Philadelphia, Pennsylvania, United States

Disclosure information not submitted.

Purpose/Background: Diverticular disease is estimated to be 40-53% heritable, and recent work has begun to define the genetic architecture of diverticulitis. A polygenic risk score (PRS) is a continuous, cumulative risk prediction tool that is derived from summing the small risk contributions of individual DNA variants and has shown promise in other diseases such as coronary artery disease. One PRS has been published for diverticulitis and it demonstrated predictive potential, but was tested only in a small population of European-ancestry individuals, raising questions about its application in larger cohorts that include non-European-ancestry participants.

Methods/Interventions: We constructed a weighted allele score of 44 independent (r2 > 0.5) DNA variants associated with diverticular disease at genome-wide significance (p < 5E-8) from a prior GWAS in the UK biobank. We then calculated the PRS for each participant in the Penn Medicine BioBank (PMBB) and analyzed the relationship between diverticulitis and PRS. Risk scores were broken into deciles, with the first decile (0-10%) containing the patients with the lowest PRS. The Fisher’s exact test was used to compare the rate of diverticulitis between decile groups. The discrimination and calibration of the PRS to predict diverticulitis risk was assessed by the calculating the area under the receiver operating characteristic curve (AUROC) and Brier score, respectively.

Results/Outcomes: Of 42,226 participants, 30,076 (71%) were of European ancestry, 11,206 (27%) were of African ancestry. 1,666 (3.9%) had diverticulitis. Patients in the top risk decile based on PRS had a higher rate of diverticulitis (5.0%) than all other patients (3.7%, p < 0.001), an effect that was particularly pronounced when comparing the top to bottom risk deciles (2.5%, p < 0.001). The prevalence of diverticulitis in each risk decile (along with its 95% CI) is given in Figure 1. On multivariable logistic regression controlling for age and sex, the PRS was significantly associated with increased odds of having diverticulitis (OR 2.26, 95% confidence interval 1.94 – 2.63, p < 0.001). A model including PRS, age, and sex had an AUROC of 66.7% (95% CI 66.1% – 67.5%) and a Brier score of 0.0365 (95% CI 0.0355 – 0.0375) (Figure 2), which, when compared to the same model without the inclusion of the PRS, demonstrated significant improvement in the AUROC (64.5% [95% CI 63.9 – 65.1]) though not the Brier score (3.65 [95% CI 0.0357 – 0.0372]).

Conclusion/Discussion: This PRS was able to identify patients at high risk for diverticulitis in a large and diverse biobank and to produce an effective model to predict diverticular disease. This is the first demonstration of the utility of a diverticulitis PRS in a cohort that includes non-European-ancestry participants.