The NIH Chronic Gvhd Global Score and the CIBMTR Risk Score Correlate Well with Failure Free Survival Following Frontline Systemic Corticosteroid Therapy for Chronic Gvhd

Introduction: While the National Institutes of Health (NIH) consensus criteria (NCC) on diagnosis and staging of cGVHD are now widely used, minimal attempts have been made to validate its ability to predict the outcomes of cGVHD treatment. Arora et al. analyzed data from the Center for International Blood and Marrow Transplant Registry (CIBMTR) and suggested that the CIBMTR risk score is a predictive factor for overall survival (OS) and non-relapse mortality (NRM). The failure free survival (FFS) has been recently proposed as an endpoint for clinical trials on treatment of cGVHD. With this study we aimed to evaluate if the cGVHD global score by the NCC and the CIBMTR risk groups (RG) could predict the treatment outcomes of frontline systemic steroid therapy for cGVHD.

Method: We retrospectively reviewed 668 consecutive patients who underwent allo-HCT between 2004 and 2012 at the Princess Margaret Cancer Centre, Toronto, Canada. We then identified 312 patients who had a diagnosis of cGVHD and received at least one systemic treatment for this. Thirty-five patients were excluded from the analysis as they were not treated with systemic corticosteroids and the remaining 277 patients were included in the final analysis. Chronic GVHD was reclassified and graded using the NCC. The CIBMTR risk score was also calculated and patients were stratified into 3 RGs accordingly. We evaluated the treatment outcomes including OS, NRM, relapse and FFS. FFS was defined as time to a switch in systemic therapy (i.e. failure of the treatment), NRM or relapse. The Kaplan-Meier method was used for OS and FFS. The cumulative incidences of NRM and relapse were calculated considering competing risks. Multivariate analysis was performed using the Cox proportional hazard regression model for OS and FFS. The Fine-Gray method was used for the incidences of NRM and relapse in multivariate analysis.

Results: With a median follow-up duration of 26 months among survivors, the median to onset of cGVHD was 140 days (range, 45-381). One hundred and two patients (36.9%) were classified as classical cGVHD and 175 patients (63.1%) as overlap syndrome. At the onset of cGVHD 90 patients (32.5%) had mild cGVHD by the NIH global score (NIH GS), moderate in 143 patients (51.6%) and severe in 44 patients (15.9%). Thirty-three patients (11.9%) presented with progressive type onset (PTO). The CIBMTR risk score was available in 227 patients and we were able to stratify them into those with RG1 (score 0-2; n=32, 14.1%), RG2 (score 3-6; n=162; 71.4%) and RG3-6 (score ≥7; n=33, 14.5%).

The median FFS of the 277 patients was 255 days. A severe GS correlated with the worst FFS: median FFS duration was 164 days in severe vs 238 days in moderate vs 304 days in mild (p=0.001). The CIBMTR RG also reflected the prognosis of patients after cGVHD treatment: the higher the RG the shorter the median duration of FFS: 166 days in RG3-6 vs 291 days in RG2 vs 501 days in RG1 (p=0.003).

The OS at 2 years from the onset of cGVHD was 74.3%. A severe GS was associated with a worse OS: 55.1% in severe vs 79.9% in moderate vs 76% in mild grade (p<0.001). The CIBMTR RG was predictive of OS: 92% in RG1 vs 81.5% in RG2 vs 38.9% in RG3-6 (p<0.001).

The cumulative incidence of NRM was 7.6% at 2 years. The NIH GS was predictive of NRM. The 2-year NRM rate was lowest with a mild GS at 2.4%, intermediate with a score of moderate at 4.8%, but significantly higher with a severe score at 27.5% (p<0.001). The CIBMTR RGs were also predictive of NRM: 0% for RG1, 6.1% for RG2 and 21.9% for RG3-6 (p=0.001). The cumulative incidence of relapse at 2 years was 9.2% and it was not associated with either the NIH GS or CIBMTR RGs..

Multivariate analysis confirmed that an NIH GS of severe and the CIBMTR RG3-6 and the overlap syndrome were associated with worse FFS: A severe NIH GS hazard ratio (HR) 1.65, p=0.011; CIBMTR RG3-6, HR 2.39, p=0.006; overlap syndrome, HR 1.38, p=0.058. A NIH GS of severe and CIBMTR RG3-6 were verified as adverse risk factors for NRM: A severe NIH GS, HR 6.31, p=0.001; CIBMTR RG3-6, HR 3.44, p=0.0081. CIBMTR RG3-6 and PTO were identified as unfavorable factors for OS: CIBMTR RG3-6, HR 5.14, p=0.002; PTO, HR 3.20, p<0.001.

Conclusion: The NIH GS and CIBMTR RG for cGVHD correlated well with FFS and NRM following first line systemic treatment for cGVHD with corticosteroid-based regimens.