Damodaran, Sindhuja Sominidi

We performed a retrospective review of patients with lichen planus pemphigoides (LPP) over almost a 23-year period to gather clinical, histopathological, immunological and treatment-related information and to compare idiopathic LPP with drug-induced LPP. This study represents the largest single-institution case series of LPP to date and includes novel findings related to mucosal involvement, autoantibody profiles and the value of monitoring bullous pemphigoid (BP) autoantibodies in predicting treatment response. This study also includes characteristic histopathological findings that can aid in the diagnosis and differentiation of LPP from BP, in addition to comparisons between idiopathic LPP and drug-induced LPP.

Abstract Acute graft-versus-host disease (GVHD), a serious complication of allogeneic hematopoietic cell transplantation (HCT), frequently involves the skin. Since clinicopathologic features of GVHD can mimic those of certain drug eruptions, accurate diagnosis can be challenging. We aimed to develop a tissue-based molecular assay to improve differentiation of GVHD from vacuolar interface drug eruption in the post-HCT period. After identifying 62 putative tissue-based biomarkers using unbiased molecular methods, we incorporated 26 into a quantitative RT-PCR assay. We tested skin biopsy specimens showing vacuolar interface changes attributable to either acute GVHD (n=67) or drug reaction (n=17), as classified by meticulous review of clinical features and course. Data were incorporated into clinical-molecular fusion diagnostic models. Significant differences in the expression of several genes, including MX1, MNDA, OAS, LIMA1, GSTM5, and SPP1, were detected between the two diagnostic groups (p<0.05). The optimal fusion model, which incorporated quantitative expression of MX1, MNDA, OAS, and GSTM5, as well as the clinical feature of diarrhea, imparted high diagnostic accuracy (receiver operating characteristic area-under-curve: 0.89). This novel tissue-based diagnostic molecular assay and fusion model demonstrates potential to improve distinction of GVHD from drug eruptions in the post-HCT setting. Multicenter validation studies are required.

e22088 Background: In recent years, adjuvant therapy trials in stage III melanoma have been successful and trials have started with the inclusion of stage IIB/C patients. However, stage IIA melanoma patients are currently not eligible for adjuvant therapy, even though a large part of all melanoma related deaths occur in this patient group. Therefore, a strong clinical need has emerged for diagnostic tools that can identify high-risk patients who currently have no access to adjuvant therapy. Here, we sought to assess the ability of a recently introduced clinicopathologic gene expression model (CP-GEP) (Bellomo et al., JCO Precis Oncol. 2020: in press) to select stage IIA patients at high risk for disease relapse, upon design of a stage-specific operating point. Methods: We assessed the prognostic performance of the CP-GEP model in all 141 stage IIA patients from a Mayo Clinic cohort of 837 consecutive melanoma patients who had a sentinel lymph node biopsy (SLNb) performed within 90 days of their diagnosis. The CP-GEP model combines Breslow thickness and patient age, with the expression of 8 genes in the primary tumor. Moreover, it stratifies patients according to their risk of relapse: CP-GEP High Risk or CP-GEP Low Risk, based on an operating point that was specifically developed for stage IIA. This stage-specific operating point was selected to fulfill the following criteria: hazard ratio RFS > 2 with a p-value < 0.05, and risk groups of similar size. The main clinical endpoint was five-year relapse free survival (RFS). Results: The CP-GEP High Risk group corresponds to 45% (63/141) of all stage IIA patients and captures 62% (18/29) of the total relapses in this substage. Moreover, CP-GEP High Risk patients relapse more frequently than CP-GEP Low Risk patients (RFS of 56% versus 78%; HR, 2.23; P < 0.05). The prognosis for stage IIA CP-GEP High Risk patients in our cohort is worse than for stage IIC/IIIA patients with reported RFS ranging from 63% to 77%. Conclusions: The CP-GEP model can be optimized by designing a stage-specific operating point, to identify a subset of stage IIA patients with an increased risk for disease relapse, not very different from IIC/IIIA patients. Therefore, stage IIA CP-GEP High Risk patients may be considered for inclusion in adjuvant trials. Independent validation studies are ongoing for the newly developed operating point. [Table: see text]

10068 Background: The identification of early stage melanoma patients at high risk for relapse is still difficult. Roughly 50% of melanoma deaths occur in patients who were initially diagnosed with nonmetastatic melanoma. Therefore, a strong clinical need has emerged for diagnostic tools that can identify melanoma patients at high risk for relapse. Here, we assessed the performance of a recently developed model (Bellomo et al., JCO Precis Oncol. 2020: in press), combining clinicopathologic and gene expression variables (CP-GEP), in identifying melanoma patients that have a high risk for disease relapse. Methods: We assessed the prognostic performance of the CP-GEP model in a cohort of 837 consecutive melanoma patients from Mayo Clinic who had a sentinel lymph node biopsy (SLNb) performed within 90 days of their diagnosis. The CP-GEP model combines Breslow thickness and patient age, with the expression of 8 genes in the primary tumor, to stratify patients according to their risk of relapse: CP-GEP High Risk or CP-GEP Low Risk. The main clinical endpoint of this study was five-year relapse free survival (RFS). Results: Patients were stratified based on SLNb status and CP-GEP classification. 76% of the patients were SLNb negative and had an RFS of 79% versus 52% for SLNb positive patients; HR, 3.21; P < 0.0001. 60% of the patients were identified as CP-GEP High Risk and had an RFS of 62% versus 87% for CP-GEP Low Risk patients; HR, 4.12; P < 0.0001. Within the SLNb negative group (637 patients of which 65% stage I), 51% of patients were classified as CP-GEP High Risk. Here, RFS was 70% for CP-GEP High Risk patients versus 89% for CP-GEP Low Risk patients; HR, 3.61; P < 0.0001. The prognosis of these CP-GEP High Risk patients is similar to stage IIC/IIIA patients with reported RFS ranging from 63% to 77%. This confirms the heterogeneity in prognosis among patients with stage I/II melanoma disease. Conclusions: The CP-GEP model can be successfully used to stratify patients based on their risk for relapse. In particular, it can be used to identify SLNb negative patients with a high risk for disease relapse who may benefit from therapeutic interventions. Independent validation studies are ongoing to validate the CP-GEP model in various patient populations. [Table: see text]