Doctor of Philosophy
Ward, Aaron D.
Palma, David A.
Lung cancer is characterized by its aggressiveness, heterogeneity, and wide array of treatments. Choosing the best treatment requires extensive patient information, which may sometimes be incomplete, presenting several clinical challenges. This thesis addresses two such challenges and introduces three machine-learning-based models to address them.
The first challenge focuses on identifying lung squamous cell carcinoma (SqCC) patients who could benefit from immunotherapy based on their tumour mutational burden (TMB). TMB is a promising but often unavailable biomarker in routine clinical practice. To address this, we developed a model capable of predicting TMB from standard-of-care tumour excision slides. Using 50 slides from 35 centres, we found that VGG16 had an area under the receiver operating characteristic curve (AUC) of 0.65, demonstrating that TMB status can be inferred from tissue morphology.
A crucial aspect of this study was the need to measure cancer tissue content within tiles from the tumour resections, which requires labour-intensive manual contours by expert pathologists. To automate this, we developed a separate VGG16-based model using 116 scans of lung SqCC tumour excisions from 35 centres. The model demonstrated a median regression error of 4% with a standard deviation of 36%, and an AUC of 0.83 at a 50% cancer content threshold. By automating this process, we can scale up TMB prediction models, making them more clinically applicable.
The second clinical challenge pertains to distinguishing benign radiation-induced lung injury (RILI) from tumour recurrence following stereotactic ablative radiotherapy (SABR). SABR is highly effective for early-stage inoperable lung cancer but often leads to RILI, which appears similar to tumour recurrence on CT scans. Accurate differentiation is critical for deciding whether invasive testing or salvage therapy is necessary. Utilizing CT scans from 68 patients showing lesion growth post-SABR, bootstrapped experiments with a random forest classifier had an average AUC of 0.66. Notably, the features deemed important by the model were correlated with clinical outcomes, marking an important advancement in non-invasive distinction of RILI from recurrence.
These studies highlight the potential of machine-learning to address critical clinical challenges in lung cancer care, particularly in the context of novel treatments like immunotherapy and SABR.
Summary for Lay Audience
Lung cancer is a complex disease with a wide array of treatment options, yet fully personalized care for it remains elusive due to the need for extensive patient information that is sometimes incomplete. This thesis addresses two clinical challenges where this is the case by using artificial intelligence to build models that predict the missing information.
The first challenge focuses on predicting tumour mutational burden (TMB), a biomarker that measures how mutated a cancer is. This biomarker is important for determining which lung cancer patients are most likely to benefit from immunotherapy, which can be highly effective but carries the risk of severe side effects. Unfortunately, TMB assessment is currently inaccessible in most clinics, as it requires invasive biopsy and costly genetic sequencing. However, this thesis shows that a model can predict TMB from existing lung tissue images based on the microscopic appearance of cancer cells on tissue samples taken during previous cancer surgeries, which are common for patients who are considered for immunotherapy. Tumours from surgery contain many non-cancer tissues, and these tissues were excluded manually when developing the model, but to improve clinical translatability, this thesis also presents a model to automate this by identifying cancer cells within surgically-removed tumours. This would allow the TMB prediction model to be completely automated.
The second challenge involves assessing the success of stereotactic ablative radiotherapy (SABR). SABR is a highly effective and well-tolerated treatment, but it induces benign inflammation and scarring around the tumour, mirroring tumour growth in post-treatment scans. Distinguishing these benign changes from true cancer recurrence is critical, as the latter requires immediate and potentially risky interventions. With current tools, this takes over a year, but this thesis presents a model that can discern the two at the earliest sign of potential growth. The model can do this by analyzing the appearance of lesions on routine follow-up scans, detecting patterns of recurrence before they become visible to the human eye.
These studies highlight the potential of artificial-intelligence-based models to address critical clinical challenges in lung cancer care, particularly in the context of novel treatments like immunotherapy and SABR.
Dammak, Salma, "Using Machine Learning Models to Address Challenges in Lung Cancer Care" (2023). Electronic Thesis and Dissertation Repository. 9787.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Available for download on Sunday, October 20, 2024