Search

Background: Genome DNA methylation profiling is a promising yet costly method for cancer classification, involving substantial data. We developed an ensemble learning model to identify cancer types using methylation profiles from a limited number of CpG sites., Methods: Analyzing methylation data from 890 samples across 10 cancer types from the TCGA database, we utilized ANOVA and Gain Ratio to select the most significant CpG sites, then employed Gradient Boosting to reduce these to just 100 sites., Results: This approach maintained high accuracy across multiple machine learning models, with classification accuracy rates between 87.7% and 93.5% for methods including Extreme Gradient Boosting, CatBoost, and Random Forest. This method effectively minimizes the number of features needed without losing performance, helping to classify primary organs and uncover subgroups within specific cancers like breast and lung., Conclusions: Using a gradient boosting feature selector shows potential for streamlining methylation-based cancer classification., (© 2024. The Author(s).)

We retrospectively evaluated long-term outcomes of high dose chemotherapy followed by autologous stem cell transplant (HDC/ASCT) in patients with diffuse large B-cell lymphoma (DLBCL). Between 2004 and 2020, 46 DLBCL patients received HDC/ASCT in our institution, including 12 patients (26.1%), who received as an upfront setting (UFS). At a median follow-up time of 69 months (range, 2-169 months), the 5-year progression-free survival (PFS) rates were 82.5% (95%CI, 46.1-95.3%) in the UFS, and 57.8% (95%CI, 38.1-73.2%) in the relapsed or refractory (R/R) patients (n=34), respectively. The 5-year PFS rates were 62.3% (95%CI, 34.0-81.3%) in primary resistant (n=13) or early relapsing (within 1 year from the initial diagnosis) patients (n=4), and 53.3% (95%CI, 25.9-74.6%) in those relapsing >1 year after the initial diagnosis (n=17), with no statistically significant difference (p=0.498). In R/R patients, multivariate analysis showed that the remission status before HDC/ASCT was an independent poor prognostic factor for progression-free survival (hazard ratio [HR], 17.0; 95%CI, 3.35-86.6; p=0.000630) and high-risk category in the international prognostic index for OS (HR, 9.39; 95%CI, 1.71-51.6; p=0.0100). The incidence of non-relapse mortality by 5 years, and 10 years were 12.2%, and 15.2%, respectively. Eleven patients (23.9%) developed second malignancies, which was the most frequent late complication after HDC/ASCT, with 5-year, and 10-year cumulative incidence of 16.9%, 22.5%, respectively. In conclusion, HDC/ASCT is effective for chemo-sensitive R/R DLBCL regardless of the timing and lines of therapy. However, careful observation is required, considering the long-term complications such as secondary malignancies.

We previously examined the utility of rituximab-bendamustine (RB) in patients with follicular lymphoma (FL) exhibiting less than optimal responses to 2 cycles of the R-CHOP chemotherapy regimen. The aim of this study was to identify molecular biomarkers that can predict prognosis in RB-treated patients in the context of the prospective cohort. We first analyzed the mutational status of 410 genes in diagnostic tumor specimens by target capture and Sanger sequencing. CREBBP, KMT2D, MEF2B, BCL2, EZH2, and CARD11 were recurrently mutated as reported before, however none was predictive for progression-free survival (PFS) in the RB-treated patients (n = 34). A gene expression analysis by nCounter including 800 genes associated with carcinogenesis and/or the immune response showed that expression levels of CD8 + T-cell markers and half of the genes regulating Th1 and Th2 responses were significantly lower in progression of disease within the 24-mo (POD24) group (n = 8) than in the no POD24 group (n = 31). Collectively, we selected 10 genes (TBX21, CXCR3, CCR4, CD8A, CD8B, GZMM, FLT3LG, CD3E, EOMES, GZMK), and generated an immune infiltration score (IIS) for predicting PFS using principal component analysis, which dichotomized the RB-treated patients into immune IIS high (n = 19) and IIS low (n = 20) groups. The 3-y PFS rate was significantly lower in the IIS low group than in the IIS high group (50.0% [95% CI: 27.1-69.2%] vs. 84.2% [95% CI: 58.7-94.6%], P = .0237). Furthermore, the IIS was correlates with absolute lymphocyte counts at diagnosis (r = 0.460, P = .00355). These results suggest that the T-cell-associated immune markers could be useful to predict prognosis in RB-treated FL patients. (UMIN:000 013 795, jRCT:051 180 181)., (© 2021 The Authors. Cancer Science published by John Wiley & Sons Australia, Ltd on behalf of Japanese Cancer Association.)