Thank you to the JMI editorial team at the 2023 RSNA for setting the pace for the fantastic collection of 2024 special sections.

As we bring the 11th volume of the Journal of Medical Imaging (JMI) to a close, I am reflecting on the remarkable progress and contributions we’ve witnessed over the past year. It has been an extraordinary privilege to work with our editorial team of over 56 researchers and spanning five continents. Together, we have formed a coherent and collaborative network, delivering timely reviews and thoughtful feedback across a dozen time zones.

One of the highlights of the year has been our special sections, which represent some of the most impactful community-driven efforts in the journal’s history. These collections have consistently demonstrated higher citation rates—by about 30%— than regular issue articles, showcasing their value in driving discovery and advancing discussions. Moving into 2025, we have two open special section calls:

We encourage prospective special section organizers to discuss their ideas with our editorial board at our focus conferences including SPIE Medical Imaging or RSNA.

Our most-cited articles from the 10th volume (2023) include “ Toward fairness in artificial intelligence for medical image analysis: identification and mitigation of potential biases in the roadmap from data collection to model deployment” (led by Karen Drukker, doi: 10.1117/1.JMI.10.6.061104) and “ CaraNet: context axial reverse attention network for segmentation of small medical objects” (led by Ange Lou, doi: 10.1117/1.JMI.10.1.014005), and our all-time highest cited article remains “ Recurrent residual U-Net for medical image segmentation” (led by Md Zahangir Alom, doi: 10.1117/1.JMI.6.1.014006). These works exemplify the timely, rigorous, and innovative spirit that defines JMI.

While celebrating successes, I feel that my year-end reflection should also address the challenges of the year. A significant portion of my duties as Editor-in-Chief involves desk rejecting submissions—currently over 50%. I posit that discussing common concerns could help our readership avoid these issues. Specifically, I offer the following synopsis of my last year of observations:

JMI focuses on medical imaging processes, contributions, and innovations. Manuscripts that emphasize clinical phenotypes or disease-focused narratives without imaging insights are better suited for clinically oriented journals.

Submissions must provide a clear quantitative data analysis. P-values are a great start but are often insufficient when viewed in isolation; effect sizes, confidence intervals, and clinical relevance should also be considered. Null findings, when well-validated, are welcome but must demonstrate clear statistical grounding such as clear characterization of confidence intervals.

I strongly value work that builds upon public datasets and challenges, but proper citations and clear comparisons to leaderboards or state-of-the-art methods are essential. Baselines should be thoughtfully selected and justified—merely applying popular frameworks like ResNet or U-Net without context does not suffice. When possible, provide clear and direct link to leaderboards from the conclusion of the challenge and/or clear descriptions of the current state-of-the-art leaderboard.

The expression “publish or perish” has been offered as the rallying cry for academic science, but a paper is not simply a stake holder of “We did this.” Manuscripts need to teach. (Side note: The root for the word “doctor” is the Latin docere, meaning “to teach.”) Whether a work introduces generalizable methods, reusable datasets, validated protocols, or solutions to unsolved problems, the manuscript should clearly define its intended audience and relevance.

Citation data offers one lens to evaluate impact. While not the only measure, articles without citations at three years post-publication may indicate a missed opportunity to address a timely or significant audience. Recently, I worked with the OpenAI 4o model and the PDFs of articles that had not been cited (according to Web of Science). We came up with the following themes in the scientific approach that may have reduced relevance for scientific follow-up:

As we turn the page to another year, I thank you, our readers, authors, reviewers, and editors, and staff, for your partnership and trust. Here’s to another year of breakthroughs and collaboration.