Machine learning predicts stem cell transplant response in severe scleroderma

Abstract

Objective: The Scleroderma: Cyclophosphamide or Transplantation (SCOT) trial demonstrated clinical benefit of haematopoietic stem cell transplant (HSCT) compared with cyclophosphamide (CYC). We mapped PBC (peripheral blood cell) samples from the SCOT clinical trial to scleroderma intrinsic subsets and tested the hypothesis that they predict long-term response to HSCT.

Methods: We analysed gene expression from PBCs of SCOT participants to identify differential treatment response. PBC gene expression data were generated from 63 SCOT participants at baseline and follow-up timepoints. Participants who completed treatment protocol were stratified by intrinsic gene expression subsets at baseline, evaluated for event-free survival (EFS) and analysed for differentially expressed genes (DEGs).

Results: Participants from the fibroproliferative subset on HSCT experienced significant improvement in EFS compared with fibroproliferative participants on CYC (p=0.0091). In contrast, EFS did not significantly differ between CYC and HSCT arms for the participants from the normal-like subset (p=0.77) or the inflammatory subset (p=0.1). At each timepoint, we observed considerably more DEGs in HSCT arm compared with CYC arm with HSCT arm showing significant changes in immune response pathways.

Conclusions: Participants from the fibroproliferative subset showed the most significant long-term benefit from HSCT compared with CYC. This study suggests that intrinsic subset stratification of patients may be used to identify patients with SSc who receive significant benefit from HSCT.

Keywords: cyclophosphamide; systemic sclerosis; treatment.

Figure 1

CONSORT diagram (1) One baseline sample failed quality control thresholds in analysis. The 8-month timepoint was initially used as a proxy for intrinsic subset assignment for this patient but was not used in the final survival analysis in presented in Figure 3. (2) A subset of subjects with available follow-up specimens was selected with preference given to those with the most available specimens at the desired timepoints. (3) The last available timepoint for subject 63 was at month 44 and was grouped with the month 48/54 samples. CONSORT, Consolidated Standards of Reporting Trials; CYC, cyclophosphamide; HSCT, haematopoietic stem cell transplant.

Figure 2

Intrinsic subsets in blood. (A) Gene expression values are displayed in a heatmap where each row is a gene in a WGCNA module and each column is a participant, ordered and labelled by intrinsic subset at the top. The eigenvalues for each sample, stratified by intrinsic subset, are plotted for the (B) inflammatory module, (C) fibroproliferative module and (D) normal-like module. P values are for Wilcoxon rank sum test. For B–D, n=20, 23 and 20 for inflammatory, normal-like and fibroproliferative intrinsic subsets. GO, gene ontology; WGCNA, weighted gene coexpression network analysis.

Figure 3

Survival analysis stratified by intrinsic molecular subset. (A) The number of participants assigned to each intrinsic subset at baseline was not significantly different between treatment arms. EFS analysis by treatment arm (HSCT in red, CYC in blue) for participants assigned to the (B) normal-like, (C) inflammatory and (D) fibroproliferative intrinsic subsets. P values are for Fisher’s exact test (A) and Mantel-Haenszel (log-rank) χ² test (B–D). CYC, cyclophosphamide; EFS, event-free survival; HSCT, haematopoietic stem cell transplant.

Figure 4

Differential gene expression analyses. Comparison between HSCT and CYC arms in terms of (A) number of available samples and (B) number of DEGs for each baseline/follow-up analysis. (C) longitudinal changes in sample functional terms in HSCT arm. The data structure was driven by clusters of earlier (8, 14, 20/26 months) and later (38, 48/54 months) time points. Each cell in a heatmap represents an ‘enrichment log’ for a given functional term at a given follow-up time point reflecting the significance of a functional enrichment (see the Methods section) in at least 2/5 comparisons. negative ‘enrichment logs’ correspond to functional terms for DEGs downregulated at follow-up and positive ‘enrichment logs’ correspond to functional terms for DEGs upregulated at follow-up with respect to baseline. (D) comparison between HSCT and CYC arms in terms of DEGs for participants with EFS and specimens available for all time points. CYC, cyclophosphamide; DEGs, differentially expressed genes; EFS, event-free survival; HSCT, haematopoietic stem cell transplant.