Page 137 - TD-4-1

Tumor Discovery

LETTER TO EDITOR
Dosiomics in lung cancer

Melek Yakar*
Department of Radiation Oncology, Faculty of Medicine, Osmangazi University, Eskişehir, Turkey

Dear Editor,
I am writing to express my opinion on dosiomics as a very current and developing topic
in lung cancer treatment.
Dosiomics has emerged as a promising tool in oncology that uses advanced
radiotherapy dose distribution features to predict clinical outcomes. In lung cancer,
where individualized treatment strategies are critical, dosiomics offers a unique approach
to integrating radiotherapy planning with predictive analysis. The focus of dosiomic
research has been to predict toxicity such as radiation pneumonitis or fibrosis. However,
recent studies have begun to explore its potential in predicting oncological outcomes
such as tumor control and survival. Despite these developments, literature pertaining to
1
this new topic remains quite limited. This letter aims to highlight the dual applications
of dosiomics in lung cancer, discuss current limitations, and suggest future directions
for research.

Inspired by radiomics, dosiomics is being developed to parameterize the dose
distribution of regions of interest using textural features and to allow the spatial
description of the dose distribution. The first studies on the use of dosiomics in lung
2-7
cancer were set to predict treatment-related toxicity. Radiation-induced toxicity
remains a major challenge in radiotherapy, especially in the treatment of lung cancer,
where critical body structures such as the lung, heart, and esophagus are in close proximity
*Corresponding author: to the tumor. Studies have shown that dosiomics can extract meaningful features from
Melek Yakar
(myakar@ogu.edu.tr) dose distribution maps, such as texture, shape, and density, which are associated with
2,3
the likelihood of developing radiation pneumonitis or esophageal toxicity. Dosiomic
Citation: Yakar M. Dosiomics
in lung cancer. Tumor Discov. features can be extracted using a handcrafted method. Handcrafted features have been
2-7
2025;4(1):129-131. used to predict toxicity such as radiation esophagitis, pneumonia, and lymphopenia.
doi: 10.36922/td.8465 Handcrafted features contain 3D dose distribution information but do not fully reflect
Received: January 9, 2025 it. If well trained, the convolutional neural network method can reveal more detailed
2
features hidden in the original data, for example, upon testing radiomics, dosiomics,
Accepted: January 14, 2025
clinical features, and dose-volume histogram, both in isolation and combination, Zhang
Published online: January 31, et al. revealed that the hybrid model achieved high prediction accuracy rates. These
7
2025
findings highlight the potential of dosiomics to improve patient selection and treatment
Copyright: © 2025 Author(s). planning, thereby reducing the incidence of severe toxicity. Furthermore, integrating
This is an Open-Access article dosiomics with clinical and radiomics data increases the predictive power, allowing for
distributed under the terms of the
Creative Commons Attribution a more comprehensive risk assessment. 7
License, permitting distribution,
and reproduction in any medium, Tumor control probability and progression-free survival are critical measures
provided the original work is for evaluating the efficacy of lung cancer treatments. Recent studies have shown that
properly cited. dosiomics can provide insights into these outcomes by analyzing the spatial and volumetric
Publisher’s Note: AccScience distribution of radiation dose within the tumor and surrounding tissues. For example,
Publishing remains neutral with Bhandari et al. conducted a study on lung cancer patients undergoing stereotactic
8
regard to jurisdictional claims in
published maps and institutional body radiotherapy (SBRT) and effectively predicted treatment failure in the lung SBRT
affiliations. treatment with a dosiomic model that integrated the interaction between computed
Volume 4 Issue 1 (2025) 129 doi: 10.36922/td.8465