Epidemiology, methodology, and biomarkers for lung cancer

<p dir="ltr">Lung cancer is the leading cause of cancer-related death and among the most diagnosed cancers worldwide. Although mortality has declined in recent decades with reduced smoking prevalence, lung cancer remains a major public health challenge. Reducing its burden requires precise risk stratification, targeted screening of high-risk populations, and robust molecular biomarkers for early detection and personalized treatment.</p><p dir="ltr">In this thesis, we integrate epidemiology, biomarker discovery, and methodology development to address the challenges of early detection, molecular stratification, risk prediction, and precision oncology. We aim to: (1) develop tools to identify clinically actionable gene fusions to resolve tumor heterogeneity and to inform targeted therapies (Study I); (2) develop advanced sequencing methods for ctDNA-based ultra-sensitive mutation detection to improve early diagnosis (Study II); (3) quantify lung cancer risk in patients with interstitial lung disease (ILD) using a sibling design, to inform screening strategies (Study III); and (4) provide a practical, reproducible protocol for SplitFusion to guide users in its implementation (Study IV).</p><p dir="ltr">In Study I, we developed SplitFusion, an ultra-sensitive algorithm for detecting clinically actionable gene fusions (e.g., ALK, ROS1, RET, and NTRK) from low- quality RNA-seq data. Leveraging the library optimization from anchored multiplex PCR (AMP) and an optimized read-splitting strategy with functional annotation and filtering, SplitFusion sensitively and accurately captures low- frequency fusions in highly degraded formalin-fixed paraffin-embedded (FFPE) samples. By resolving fusion-driven tumor heterogeneity, SplitFusion improves the molecular stratification of patients to guide targeted therapies.</p><p dir="ltr">In Study II, we developed BLAST-Seq (Bi-Strand Linear Amplification and Single-strand Target enrichment) for non-invasive early cancer diagnosis. The method enables ultra-sensitive detection of circulating tumor DNA (ctDNA) by a novel library design, with a custom computational error-suppression algorithm, incorporating UMI-based consensus calling and stringent filtering. BLAST-Seq achieved high analytical sensitivity compared with orthogonal ddPCR validation, and reliably identified driver mutations from low-input materials and at low variant allele frequencies (VAFs), enhancing early detection and minimal residual disease (MRD) monitoring.</p><p dir="ltr">In Study III, we evaluated the association between interstitial lung disease (ILD) and lung cancer to refine screening strategies for high-risk populations. Using the Swedish nationwide registers as a large population-based cohort and a sibling-controlled design, we demonstrate that ILD significantly increases the risk of incident lung cancer, among all major histological subtypes. Thereby we define a critical smoking-independent high-risk group that may benefit from enhanced surveillance.</p><p dir="ltr">In Study IV, we established a standardized and reproducible protocol for the implementation of the SplitFusion pipeline. This protocol provides comprehensive, step-by-step guidance on the required software environment, input data formats, and specific command-line instructions. It aims to offer a user-friendly framework that facilitates the consistent and reliable application of SplitFusion in both research and translational settings.</p><p dir="ltr">In summary, this thesis advances precision oncology in lung cancer by integrating epidemiological risk identification with cutting-edge molecular diagnostics. By combining ultra-sensitive mutation detection and fusion profiling, this work provides a comprehensive framework to improve prevention, early detection, and individualized therapy.</p><h3>List of scientific papers</h3><p dir="ltr">I. <b>Weiwei Bian</b>*, Baifeng Zhang*, Zhengbo Song*, Binyamin A Knisbacher*, Yee Man Chan, Chloe Bao, Chunwei Xu, Wenxian Wang, Athena Hoi Yee Chu, Chenyu Lu, Hongxian Wang, Siyu Bao, Zhenyu Gong, Hoi Yee Keung, Zi-Ying Maggie Chow, Yiping Zhang, Wah Cheuk, Gad Getz, Valentina Nardi, Mengsu Yang, William Chi Shing Cho, Jian Wang, Juxiang Chen, Zongli Zheng. SplitFusion enables ultrasensitive gene fusion detection and reveals fusion variant- associated tumor heterogeneity. Patterns, Volume 6, Issue 2, 101174. <a href="https://doi.org/10.1016/j.patter.2025.101174" rel="noreferrer" target="_blank">https://doi.org/10.1016/j.patter.2025.101174</a></p><p dir="ltr">II. Firaol Tamiru Kebede*, <b>Weiwei Bian</b>*, Chenyu Lu, Siyu Bao, Weimin Ye, William Chi Shing Cho, Zongli Zheng. Sensitive Cancer Detection Using Circulating Cell-Free DNA. [Manuscript]</p><p dir="ltr">III. Hui Xu, Li Yin, <b>Weiwei Bian</b>, Mingqiang Kang, Hans-Olov Adami, Weimin Ye. Interstitial Lung Disease and Risk of Lung Cancer. JAMA Netw Open. 2025;8(7):e2519630. <a href="https://doi.org/10.1001/jamanetworkopen.2025.19630" rel="noreferrer" target="_blank">https://doi.org/10.1001/jamanetworkopen.2025.19630</a></p><p dir="ltr">IV. <b>Weiwei Bian</b>, Binyamin A. Knisbacher, Yee Man Chan, Chen Zhao, Zongli Zheng. Protocol for sensitive detection of gene fusions from RNA-seq data using SplitFusion. [Manuscript]</p><p dir="ltr">* Equally contributed</p>