A clinical decision support tool may help to optimise vedolizumab therapy in Crohn's disease

Abstract

Background: A clinical decision support tool (CDST) has been validated for predicting treatment effectiveness of vedolizumab (VDZ) in Crohn's disease.

Aim: To assess the utility of this CDST for predicting exposure-efficacy and disease outcomes.

Methods: Using data from three independent datasets (GEMINI, GETAID and VICTORY), we assessed clinical remission rates and measured VDZ exposure, rapidity of onset of action, response to dose optimisation and progression to surgery by CDST-defined response groups (low, intermediate and high).

Results: A linear relationship existed between CDST-defined groups, measured VDZ exposure, rapidity of onset of action and efficacy in GEMINI through week 52 (P < 0.001 at all time points across three CDST-defined groups). In GETAID, CDST predicted differences in clinical remission at week 14 (AUC = 0.68) and rapidity of onset of action (P = 0.04) between probability groups. The high-probability patients did not benefit from shortening of infusion intervals, and differences in onset of action between the high-intermediate and low-probability groups within GETAID were no longer significant when including low-probability patients who received a week 10 infusion. CDST predicted a twofold increase in surgery risk over 12 months of VDZ therapy among low- to intermediate-probability vs high-probability patients (adjusted HR 2.06, 95% CI 1.33-3.21).

Conclusions: We further extended the clinical utility of a previously validated VDZ CDST, which accurately predicts at baseline exposure-efficacy relationships and rapidity of onset of action and could be used to help identify patients who would most benefit from interval shortening and those most likely to require surgery while on active therapy.

Figure 1

GEMINI 2 clinical trial 52‐week vedolizumab serum drug concentrations stratified by CDST. ^aAll values in table are median VDZ concentration (µg/mL) (IQR); post‐dose concentration was measured 2 h after dosing. ^bLow probability; ≤13 points in CDST model at baseline. ^cIntermediate probability; >13 to ≤19 points in CDST model at baseline. ^dHigh probability; >19 points in CDST model at baseline. ****P < 0.0001, ***P < 0.001, **P < 0.01, and *P < 0.05. Bolded P values are statistically significant. CDST, clinical decision support tool; IQR, interquartile range; PK, pharmacokinetics; VDZ, vedolizumab

Figure 2

GEMINI 2 clinical trial 52‐week reduction in Harvey‐Bradshaw Index stratified by CDST ^aAll values in table are mean HBI (SE). ^bLow probability; ≤13 points in CDST model at baseline. ^cIntermediate probability; >13 to ≤19 points in CDST model at baseline. ^dHigh probability; >19 points in CDST model at baseline. ****P < 0.0001, ***P < 0.001, and **P < 0.01. Bolded P values are statistically significant. CDST, clinical decision support tool; HBI, Harvey‐Bradshaw Index; LS, least‐squares; SE, standard error; Wk, week

Figure 3

GETAID vedolizumab cohort treatment outcomes stratified by CDST. (A) Treatment outcomes stratified by CDST in overall GETAID cohort. (B) Treatment outcomes stratified by CDST in GETAID cohort on Q8 week vedolizumab maintenance. (C) Reduction in HBI stratified by CDST. (D) Reduction in HBI after vedolizumab interval shortening (escalation) stratified by CDST. Low probability; ≤13 points in CDST model at baseline. Intermediate probability; >13 to ≤19 in CDST model at baseline. High probability; >19 points in CDST model at baseline. *P < 0.05. Abbreviations: CDST, clinical decision support tool; HBI, Harvey‐Bradshaw index; Q8, every 8 weeks; REM, remission; SF‐REM, steroid‐free remission. High‐probability group (n = 3); intermediate‐probability group (n = 24); low‐probability group (n = 28). Because of the small sample size of the high‐probability group (a highly refractory population early in the period during which vedolizumab became available), it was combined with the intermediate‐probability group for analyses

Figure 4

VICTORY consortium vedolizumab‐treated rates of endoscopic remission and progression to surgery stratified by CDST. A, Cumulative rates of endoscopic remission. B, Cumulative rates of progression to surgery. Low probability; ≤13 points in CDST model at baseline. Intermediate probability; >13 to ≤19 in CDST model at baseline. High probability; >19 points in CDST model at baseline. High‐probability group (n = 131), intermediate‐probability group (n = 281), low‐probability group (n = 89). Abbreviation: CDST, clinical decision support tool. Analysis of endoscopic remission limited to those patients with follow‐up endoscopic assessments (n = 326; high probability n = 84; intermediate probability n = 172; low probability n = 70). Endoscopic remission defined as absence of ulcerations. Pairwise log‐rank comparisons across the three probability groups for endoscopic remission: high vs low P < 0.001; high vs intermediate P = 0.076; low vs intermediate P = 0.002. Pairwise log‐rank comparisons across the three probability groups for progression to surgery: high vs low P = 0.024; high vs intermediate P = 0.076; low vs intermediate P = 0.264