Keywords
Neoplasms - informatics - health information technology, clinical implementation
1. Introduction
Cancer informatics (CI) is a broad field with several fundamental goals: 1) organizing
data in ways that are comprehensible and meaningful to clinicians, researchers, and
patients; 2) using data to advance the treatment of cancer; and 3) manipulating data
to yield new insights. In this edition of the Cancer Informatics section, the focus
has shifted from the increasing important area of clinical implementation. Whereas
previous efforts in cancer informatics have been largely focused on the development
of new methodologies to model cancer data, recently there has been an emphasis on
translating such methodologies within clinical practice. Both papers selected for
the 2024 describe informatics tools that either have been rigorously validated in
the context of a randomized clinical trial or leverage data that is relatively scalable
across clinical settings.
2. Paper Selection Method
One electronic database was searched: PubMed/MEDLINE. The search was performed in
January 2024 to identify peer-reviewed journal articles published in 2023, in the
English language, related to cancer informatics research. The following search was
implemented:
((“Neoplasms”[Mesh] OR “chemotherapy”) AND (“Informatics”[Mesh] OR “cancer informatics”
OR “ontologies” OR “machine learning” OR “artificial intelligence”) AND (hasabstract[text]
AND (“2023/01/01”[PDAT] : “2023/12/31”[PDAT]) AND English[lang])) NOT (“Radiotherapy
Planning, Computer-Assisted”[Mesh]) NOT (“Radiotherapy, Computer-Assisted”[Mesh])
This is identical to the search for papers in 2023 except that we also added terms
related to artificial intelligence and machine learning. This search yielded 6,366
results. Next, we excluded review articles, resulting in 1,826 articles for first-pass
review. The titles of these articles were blindly screened for relevance, resulting
in 270 articles that were reviewed in further depth. The abstract of each of the 270
was blindly reviewed and assigned as potential candidate (n=35) and non-candidate
(n=235). Given that the number of potential candidates exceeded ten, these articles
were further evaluated to select twelve final candidates.
In accordance with the IMIA Yearbook selection process, the 12 candidate best papers
were evaluated by the two section editors, senior editors, and by additional external
reviewers (at least four reviewers per paper) [[1]]. The geographic distribution of the reviewers was: eight United States, two United
Kingdom, one France, and one Denmark. Two papers were finally selected as best papers
([Table 1]). A content summary of the selected best papers can be found in the appendix of
this synopsis.
Table 1.
Selection of best papers for the 2024 IMIA Yearbook of Medical Informatics for the
section Clinical Research Informatics. The articles are listed in alphabetical order
of the first author's surname.
|
Section
Cancer Informatics
|
|
Manz C et al. Long-term Effect of Machine Learning-Triggered Behavioral Nudges on
Serious Illness Conversations and End-of-Life Outcomes Among Patients With Cancer:
A Randomized Clinical Trial. JAMA Oncology. 2023 Mar 1;9(3):414-418. doi: 10.1001/jamaoncol.2022.6303.
Placido D et al. A Deep Learning Algorithm To Predict Risk Of Pancreatic Cancer From
Disease Trajectories. Nature Medicine. 2023 May;29(5):1113-1122. doi: 10.1038/s41591-023-02332-5.
|
3. Outlook
The 12 candidate best papers for 2024 illustrate recent efforts towards data-driven
research and innovation and exemplify a diversity of different data streams across
cancer informatics. Studies analyzed data ranging from administrative codes, structured
electronic health record data, diagnostic images, and pathology imaging. The two selected
best papers cover two different areas of the oncology care spectrum, namely cancer
screening and end-of-life care. The diversity in papers highlights the maturation
of cancer informatics as a subfield and the wealth of opportunities to leverage oncology
data to personalize care.
Manz et al. [[2]] conduct a randomized clinical trial to evaluate whether a machine learning algorithm
which predicts mortality could be combined with behavioral interventions to improve
end-of-life care for patients with advanced cancer. In a trial that spanned nine clinical
practices, the authors found that clinicians exposed to the machine learning-driven
behavioral intervention were three times more likely to have early goals of care conversations
and less likely to use end-of-life systemic therapy.
Placido et al. [[3]] developed longitudinal machine learning methods based on diagnosis codes in routine
administrative data, trained and validated on 6 million patients (24,000 pancreatic
cancer cases) in Denmark and 3 million patients (3,900 cases) in the United States
(US Veterans Health Administration). Their models predicted pancreatic cancer occurrence
within incremental time windows. For the Danish cohort, the area under the receiver
operating characteristic (AUROC) curve was 0.88, with an estimated relative risk of
59 for 1,000 highest-risk patients older than age 50 years. For the US Veteran cohort,
the best-performing AUCROC=0.78.
The other 10 candidate papers cover a variety of different aspects of cancer informatics
and we outline those papers below. These papers highlight clinical model infrastructure
in cancer research, validation of machine learning models for cancer, artificial intelligence
(AI) models for diagnosis, and a special paper on federated learning. With new tools
particularly in AI being developed, the cancer informatics field continues to be very
exciting.
Building model infrastructure
Integrative Imaging Informatics for Cancer Research: Workflow Automation for Neuro-Oncology
(3CR-WANO). Journal of Clinical Oncology Clinical Cancer Informatics, 2023 [[4]].
Precision Oncology Core Data Model to Support Clinical Genomics Decision Making. Journal
of Clinical Oncology Clinical Cancer Informatics, 2023 [[5]].
The Childhood Cancer Data Initiative: Using the Power of Data to Learn from and Improve
Outcomes for Every Child and Young Adult with Pediatric Cancer, Journal of Clinical
Oncology, 2023 [[6]].
Prospective validation of machine learning models
Prospective implementation of AI-assisted screen reading to improve early detection
of breast cancer. Nature Medicine, 2023 [[7]].
Screening for extranodal extension in HPV-associated oropharyngeal carcinoma: Evaluation
of a CT-based deep learning algorithm in patient data from a multicenter, randomized
de-escalation trial. Lancet Digital Health, 2023 [[8]].
AI models focused on diagnosis of malignancy
Artificial-intelligence-based molecular classification of diffuse gliomas using rapid,
label-free optimal imaging. Nature Medicine, 2023 [[9]].
A reinforcement learning model for AI-based decision support in skin cancer. Nature
Medicine, 2023 [[10]].
Federated Learning
Federated Learning for Predicting Histological Response to Neoadjuvant Chemotherapy
in Triple-Negative Breast Cancer. Nature Medicine, 2023 [[11]].
4. Conclusion
The best papers for cancer informatics present promising frameworks and applications
for novel informatics infrastructure and artificial intelligence applications in cancer
care. We were pleased to see an emphasis on prospective applications of artificial
intelligence beginning to be represented.
