Acute Gvhd Diagnosis and Adjudication in a Multicenter Trial – a Report from the BMT CTN 1202 Biorepository Study

Abstract

Accurate and reproducible methods to diagnose, grade and report acute graft vs. host disease (GVHD) are critical for evaluation of new treatments and biomarkers. BMTCTN 1202 created a biorepository and clinical data bank for the study of GVHD biomarkers and other transplant outcomes. To create a representative study population any pt >10 kg undergoing alloHCT was eligible (Table 1). Symptoms suggestive of GVHD, biopsy results, immunosuppressive medications, and possible etiologies were reported weekly until day 100. An end-point review committee (ERC) performed a near real time adjudication of reported symptoms at their onset and assigned a confidence level (Confirmed, Probable, Possible or Negative) for each organ according to Harris, BBMT, 2016. The final repository included 41,468 annotated biospecimens from 1709 alloHCT patients (pts) transplanted at U.S. centers from 2012-2016. Although 90% of pts developed symptoms at least once, physicians suspected GVHD in only 23% of cases and GVHD was considered the only etiology in 12%. Biopsies (bx) were performed in 40% of pts, most often for upper or lower GI symptoms. Bx results were equivocal in 11% although the range among centers was broad (0-45%). Equivocal results were most likely for skin bx and during the 2nd week after transplant. Systemic steroids were administered to 70% of pts at least once, often for non-GVHD reasons (41% of steroid courses). When GVHD was in the differential diagnosis, systemic steroids were started 81% of the time. The ERC adjudicated 2,008 cases with symptoms suggestive of GVHD onset. In 12.3% of cases the ERC modified either the center reported diagnosis of GVHD (4.6%) or the automated generated severity grade (7.7%). GVHD was diagnosed in 1025 pts and categorized as confirmed (31%), probable (50%), or possible (19%). The proportion of definitive diagnoses (Confirmed or Negative) increased with longer time post-transplant (Fig 1). ERC adjudication substantially lowered the incidence of GVHD grade 2-4 from 30% as reported by centers to 23% (confirmed or probable) by ERC. An adjudicated diagnosis of GVHD strongly associated with 6 month NRM, adjusted for patient age, gender and HLA matching (HR = 2.0, p In summary, the largest prospective observational study of GVHD outcomes demonstrated that symptoms in GVHD target organs are almost universal after transplant, their differential diagnosis is broad, tissue biopsies are infrequently done and their results are often equivocal, and steroids are often prescribed for reasons other than acute GVHD. These limitations undermine the current methods of diagnosis and reporting of acute GVHD on clinical trials. A robust reporting and adjudication system has significant value and may improve the chances of success of prospective clinical trials.