Saved in:
| Main Authors: | , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2604.20869 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866913056304398336 |
|---|---|
| author | Spiess, Philippe E. Zitu, Md Muntasir Walker, Alison Anaya, Daniel A. Wenham, Robert M. Vogelbaum, Michael Grass, Daniel Jaffer, Ali-Musa Sarnaik, Amod McMullen, Caitlin Sam, Christine Kiluk, John V. Liu, Tianshi Biachi, Tiago Powsang, Julio Chern, Jing-Yi Li, Roger Felder, Seth Reynolds, Samuel Shafique, Michael Sheehan, Alison Layman, Ashley Warfield, Cydney A. Legoas, Derrick Parrinello, Jaclyn Schmitz, Jena Eaton, Kevin Honor, Mark Felipe, Luis ElNaqa, Issam Delgado, Elier Berler, Talia Phillips, Rachael V. Francisque, Frantz Fernandez, Carlos Garcia Valdes, Gilmer |
| author_facet | Spiess, Philippe E. Zitu, Md Muntasir Walker, Alison Anaya, Daniel A. Wenham, Robert M. Vogelbaum, Michael Grass, Daniel Jaffer, Ali-Musa Sarnaik, Amod McMullen, Caitlin Sam, Christine Kiluk, John V. Liu, Tianshi Biachi, Tiago Powsang, Julio Chern, Jing-Yi Li, Roger Felder, Seth Reynolds, Samuel Shafique, Michael Sheehan, Alison Layman, Ashley Warfield, Cydney A. Legoas, Derrick Parrinello, Jaclyn Schmitz, Jena Eaton, Kevin Honor, Mark Felipe, Luis ElNaqa, Issam Delgado, Elier Berler, Talia Phillips, Rachael V. Francisque, Frantz Fernandez, Carlos Garcia Valdes, Gilmer |
| contents | Background: More than 80% of U.S. cancer care is delivered in community settings, where survival remains worse than at academic centers. Clinicians must integrate genomics, staging, radiology, pathology, and changing guidelines, creating cognitive burden. We evaluated OncoBrain, an AI clinical reasoning platform for oncology treatment-plan generation, as an early step toward OGI.
Methods: OncoBrain combines general-purpose LLMs with a cancer-specific graph retrieval-augmented generation layer, a gold-standard treatment-plan corpus as long-term memory, and a model-agnostic safety layer (CHECK) for hallucination detection and suppression. We evaluated clinician-enriched case summaries across gynecologic, genitourinary, neuro-oncology, gastrointestinal/hepatobiliary, and hematologic malignancies. Three clinician groups completed structured evaluations of 173 cases using a common 16-item instrument: subspecialist oncologists reviewed 50 cases, physician reviewers 78, and advanced practice providers 45.
Results: Ratings were highest for scientific accuracy, evidence support, and safety, with lower but favorable scores for workflow integration and time savings. On a 5-point scale, mean alignment with evidence and guidelines was 4.60, 4.56, and 4.70 across subspecialists, physician reviewers, and advanced practice providers. Mean scores for absence of safety or misinformation concerns were 4.80, 4.40, and 4.60. Workflow integration averaged 4.50, 3.94, and 4.00; perceived time savings averaged 5.00, 3.89, and 3.60.
Conclusions: In this multi-specialty vignette-based evaluation, OncoBrain generated oncology treatment plans judged guideline-concordant, clinically acceptable, and easy to supervise. These findings support the potential of a carefully engineered AI reasoning platform to assist oncology treatment planning and justify prospective real-world evaluation in community settings. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_20869 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Clinical Reasoning AI for Oncology Treatment Planning: A Multi-Specialty Case-Based Evaluation Spiess, Philippe E. Zitu, Md Muntasir Walker, Alison Anaya, Daniel A. Wenham, Robert M. Vogelbaum, Michael Grass, Daniel Jaffer, Ali-Musa Sarnaik, Amod McMullen, Caitlin Sam, Christine Kiluk, John V. Liu, Tianshi Biachi, Tiago Powsang, Julio Chern, Jing-Yi Li, Roger Felder, Seth Reynolds, Samuel Shafique, Michael Sheehan, Alison Layman, Ashley Warfield, Cydney A. Legoas, Derrick Parrinello, Jaclyn Schmitz, Jena Eaton, Kevin Honor, Mark Felipe, Luis ElNaqa, Issam Delgado, Elier Berler, Talia Phillips, Rachael V. Francisque, Frantz Fernandez, Carlos Garcia Valdes, Gilmer Computers and Society Artificial Intelligence Human-Computer Interaction Information Retrieval Machine Learning Background: More than 80% of U.S. cancer care is delivered in community settings, where survival remains worse than at academic centers. Clinicians must integrate genomics, staging, radiology, pathology, and changing guidelines, creating cognitive burden. We evaluated OncoBrain, an AI clinical reasoning platform for oncology treatment-plan generation, as an early step toward OGI. Methods: OncoBrain combines general-purpose LLMs with a cancer-specific graph retrieval-augmented generation layer, a gold-standard treatment-plan corpus as long-term memory, and a model-agnostic safety layer (CHECK) for hallucination detection and suppression. We evaluated clinician-enriched case summaries across gynecologic, genitourinary, neuro-oncology, gastrointestinal/hepatobiliary, and hematologic malignancies. Three clinician groups completed structured evaluations of 173 cases using a common 16-item instrument: subspecialist oncologists reviewed 50 cases, physician reviewers 78, and advanced practice providers 45. Results: Ratings were highest for scientific accuracy, evidence support, and safety, with lower but favorable scores for workflow integration and time savings. On a 5-point scale, mean alignment with evidence and guidelines was 4.60, 4.56, and 4.70 across subspecialists, physician reviewers, and advanced practice providers. Mean scores for absence of safety or misinformation concerns were 4.80, 4.40, and 4.60. Workflow integration averaged 4.50, 3.94, and 4.00; perceived time savings averaged 5.00, 3.89, and 3.60. Conclusions: In this multi-specialty vignette-based evaluation, OncoBrain generated oncology treatment plans judged guideline-concordant, clinically acceptable, and easy to supervise. These findings support the potential of a carefully engineered AI reasoning platform to assist oncology treatment planning and justify prospective real-world evaluation in community settings. |
| title | Clinical Reasoning AI for Oncology Treatment Planning: A Multi-Specialty Case-Based Evaluation |
| topic | Computers and Society Artificial Intelligence Human-Computer Interaction Information Retrieval Machine Learning |
| url | https://arxiv.org/abs/2604.20869 |