2012P - Spatially resolved transcriptomics deciphers inter- and intra-tumor heterogeneity of small cell lung cancer

Background

Small cell lung cancer (SCLC) is a highly heterogeneous cancer. Recent studies have illustrated that inter- and intra-tumor heterogeneity (ITH) is correlated with biological behavior and therapeutic vulnerability in multiple malignancies. However, it remains unclear in SCLC. Here, we seek to define ITH at the spatial transcriptome level and elucidate the crosstalk between ITH and tumor immune microenvironment (TIME) with the clinical characteristics of SCLC.

Methods

We performed digital spatial profiling (DSP) to quantify the whole transcriptomes and calculated the ITH score using the Deviating gene Expression Profiling Tumor Heterogeneity algorithm (DEPARTMENTH) at the RNA level. Deconvoluted algorithms were adopted to calculate immune infiltration from RNA expression data, which was validated with immunochemistry (IHC). A machine learning model (THIM) was constructed to determine the tumor heterogeneity and facilitate prognostic stratification, which were validated across multi-center cohorts.

Results

Using DSP RNA data of the tumor core area (79 regions) from resected limited stage SCLC (25 patients), we differentiated intra-/inter-tumor heterogeneity subtypes. SCLC tumors could be classified into two distinct subtypes: high heterogeneity+complex (HC) and medium+low heterogeneity (ML), with significant variations in CD8+ T cell infiltration and multiple biological properties. There were significant differences in overall survival (log-rank p = 0.046) and disease-free survival (log-rank p=0.024) between the two subtypes, where MLs had good and HCs showed poor prognosis. Furthermore, we developed and verified a robust THIM model using machine learning techniques which is able to predict the therapeutic response (AUC=0.923) and survival of SCLC patients, as demonstrated in both in-house and multicenter bulk transcriptomic data cohorts.

Conclusions

In summary, we have deciphered inter- and intra-tumor heterogeneity of SCLC and proposed a new heterogeneity-oriented subtyping framework for improving clinically relevant risk stratification, which is promising for postoperative stratified management of limited stage SCLC patients as well as assessment of immunotherapy efficacy in advanced stage patients.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

The authors.

Funding

Has not received any funding.

Disclosure

All authors have declared no conflicts of interest.