Abstract:Early screening for cancer has proven to improve the survival rate and spare patients from intensive and costly treatments due to late diagnosis. Cancer screening in the healthy population involves an initial risk stratification step to determine the screening method and frequency, primarily to optimize resource allocation by targeting screening towards individuals who draw most benefit. For most screening programs, age and clinical risk factors such as family history are part of the initial risk stratification algorithm. In this paper, we focus on developing a blood marker-based risk stratification approach, which could be used to identify patients with elevated cancer risk to be encouraged for taking a diagnostic test or participate in a screening program. We demonstrate that the combination of simple, widely available blood tests, such as complete blood count and complete metabolic panel, could potentially be used to identify patients at risk for colorectal, liver, and lung cancers with areas under the ROC curve of 0.76, 0.85, 0.78, respectively. Furthermore, we hypothesize that such an approach could not only be used as pre-screening risk assessment for individuals but also as population health management tool, for example to better interrogate the cancer risk in certain sub-populations.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is: how to use routine laboratory test indicators (such as complete blood count and comprehensive metabolic panel) to identify patients with an increased risk of cancer, in order to encourage them to undergo diagnostic tests or participate in screening programs. Specifically, the author hopes to use a deep - learning model to identify high - risk groups who may have colorectal cancer, liver cancer and lung cancer in advance based on common blood biomarkers, thereby increasing the early detection rate of these cancers, improving the survival rate of patients, and reducing the high treatment costs due to late - stage diagnosis. ### Main problem decomposition 1. **The need for early cancer screening**: - Early screening has been proven to improve the survival rate of cancer patients and avoid expensive and complicated treatments due to late - stage diagnosis. - Screening procedures usually involve an initial risk stratification step to determine the screening method and frequency, optimize resource allocation, and ensure that the screening targets the most benefited population. 2. **Limitations of existing screening methods**: - At present, most screening procedures rely on age and clinical risk factors (such as family history), but these factors cannot comprehensively cover all potential high - risk individuals. - The participation rate in cancer screening is low, partly because the general public has insufficient awareness of their own risks. 3. **Proposing a new risk assessment tool**: - The paper proposes a risk stratification method based on blood biomarkers, aiming to identify individuals with a higher risk of cancer, so as to encourage them to undergo further diagnosis or screening. - By combining simple and widely available blood tests (such as complete blood count and comprehensive metabolic panel), this method is expected to be used to identify high - risk patients with colorectal cancer, liver cancer and lung cancer. ### Method overview - **Data source**: The study used data of patients diagnosed with malignant tumors (including colorectal cancer, liver cancer and lung cancer) or without a malignant tumor diagnosis between 2017 and 2021. - **Model development**: The author developed a deep - learning model named Deep Profiler, which predicts the probability of a patient developing cancer within the next 12 months based on age, gender and routine blood biomarkers (such as various indicators in CBC and CMP). - **Performance evaluation**: The performance of the model on three cancers (colorectal cancer, liver cancer and lung cancer) was evaluated through the validation cohort, and the results showed that the area under the receiver operating characteristic curve (AUC) was 0.76, 0.85 and 0.78 respectively. ### Conclusion The paper shows that by using routine blood test indicators and deep - learning technology, individuals with a higher risk of cancer can be identified at an early stage, thus providing earlier intervention opportunities for these patients. This method is not only helpful for risk assessment at the individual level, but can also be used as a group health management tool to help better understand the cancer risks of certain subgroups.

Deep learning-based identification of patients at increased risk of cancer using routine laboratory markers

Towards radiologist-level cancer risk assessment in CT lung screening using deep learning

Pre-screening breast cancer with machine learning and deep learning

Lung cancer screening with low-dose CT scans using a deep learning approach

Enhancing cancer prediction in challenging screen-detected incident lung nodules using time-series deep learning

Deep Learning for Breast Cancer Risk Prediction: Application to a Large Representative UK Screening Cohort

Development and deployment of a histopathology-based deep learning algorithm for patient prescreening in a clinical trial

Deploying blood-based cancer screening

An intelligent machine learning model for real-time early detection of undesirable cancer events: AIM2REDUCE.

Deep Learning-Based, Multiclass Approach to Cancer Classification on Liquid Biopsy Data

End-to-end lung cancer screening with three-dimensional deep learning on low-dose chest computed tomography

Deep learning predicts cardiovascular disease risks from lung cancer screening low dose computed tomography

Machine Learning‐Assisted Evaluation of Circulating DNA Quantitative Analysis for Cancer Screening

Machine Learning for Early Discrimination Between Lung Cancer and Benign Nodules Using Routine Clinical and Laboratory Data

An Innovative Method for Lung Cancer Identification Using Machine Learning Algorithms

Lung Cancer Recurrence Risk Prediction through Integrated Deep Learning Evaluation

Machine learning predicts liver cancer risk from routine clinical data: a large population-based multicentric study

Early identification of patients at risk for iron-deficiency anemia using deep learning techniques

Machine Learning and Real-World Data to Predict Lung Cancer Risk in Routine Care

Emerging Strategies in Lung Cancer Screening: Blood and Beyond

Shifting the Cancer Screening Paradigm: Developing a Multi-Biomarker Class Approach to Multi-Cancer Early Detection Testing