AI Digital Twin Rivals Radiologists in Breast Cancer Surgical Planning

CHICAGO, IL – March 05, 2026 – Artificial intelligence is achieving a new level of parity with human experts in the high-stakes field of oncology. New clinical data reveals that an AI-powered “digital twin” platform can identify and measure early-stage breast cancer tumors with an accuracy comparable to that of highly trained radiologists, a development that promises to enhance precision and consistency in surgical planning.

The findings, presented this week at the 43rd Annual Miami Breast Cancer Conference, spotlight the capabilities of TumorSight® Viz, a platform developed by the AI-driven precision medicine company SimBioSys. A multi-institutional study demonstrated that the software, which transforms standard breast MRI scans into interactive 3D models, delivers reproducible anatomical insights that fall within the accepted range of variability among board-certified breast radiologists.

This marks a significant milestone in the quest to integrate AI into clinical workflows, not as a replacement for human expertise, but as a powerful tool to augment it. “This body of clinical evidence represents an important milestone in validating how AI-generated digital twins can deliver reproducible and radiologist-comparable insight into tumor anatomy,” said Stacey Stevens, President and CEO of SimBioSys, in a statement accompanying the announcement.

The AI Co-Pilot for Surgeons

At the heart of the technology is the concept of a patient-specific digital twin. TumorSight® Viz ingests standard breast MRI data and, within minutes, generates a detailed 3D visualization of the patient's unique anatomy. This model maps out the tumor's size and location in relation to critical structures like the skin, chest wall, and surrounding tissue. For surgeons, this transforms a series of flat, black-and-white images into an intuitive, explorable 3D map.

The study presented in Miami validated the platform's quantitative prowess. Key among the findings was a high spatial concordance, measured by a 0.92 surface Dice score. In medical imaging analysis, a Dice score of 1.0 represents a perfect overlap, making 0.92 an indicator of exceptionally strong agreement between the AI's automated tumor delineation and the “ground truth” measurements confirmed by expert radiologists.

Furthermore, the AI’s performance remained robust and consistent regardless of the MRI machine manufacturer, magnetic field strength, tumor subtype, or the clinical site where the imaging was performed. This level of generalizability is critical for any AI tool intended for widespread clinical adoption, as it demonstrates reliability across real-world, variable conditions.

The platform's ability to deliver these complex measurements in minutes is a significant leap in efficiency. It provides objective, anatomy-specific data that complements expert clinical interpretation, a crucial element for planning breast-conserving surgery or deciding on a mastectomy.

“Accurate visualization and quantitation of tumor extent are fundamental to surgical decision-making,” noted Dr. Barry Rosen, a breast surgical oncologist and the Chief Medical Officer of SimBioSys. “Technologies that can reliably mirror radiologist assessments have the potential to improve clarity, reproducibility, and communication across care teams.”

Validating AI in a Crowded Field

SimBioSys is not alone in applying AI to oncology, but its focus on creating a comprehensive 3D surgical planning tool provides a key differentiator. While companies like Google Health and Lunit have developed powerful AI for detecting cancer in mammograms, and others like Perimeter Medical Imaging focus on real-time margin assessment during surgery, TumorSight Viz occupies a critical niche in preoperative strategy.

The platform's credibility is bolstered by a history of regulatory validation. It has secured its third 510(k) clearance from the U.S. Food and Drug Administration (FDA), an iterative process that demonstrates the device is safe, effective, and substantially equivalent to existing tools on the market. The latest clearance for version 1.3 expanded its capabilities and improved workflow integration, including connectivity with hospital picture archiving and communication systems (PACS) to automate image transfer.

This regulatory progress, combined with a growing portfolio of peer-reviewed clinical evidence, is crucial for building trust within the medical community. For many clinicians, the “black box” nature of some AI algorithms is a significant barrier to adoption. By publishing performance data and achieving FDA clearance, the company aims to provide the transparency and validation that healthcare providers demand before integrating new technology into patient care.

Augmenting, Not Replacing, the Expert

The narrative surrounding AI in medicine often swings between utopian promises of automated perfection and dystopian fears of physician replacement. The reality, as illustrated by tools like TumorSight Viz, is far more nuanced. The prevailing view among medical professionals is that AI’s most valuable role is that of a decision-support system—a tireless assistant that can analyze vast amounts of data, identify patterns, and reduce cognitive load.

Radiologists and surgeons face immense pressure and a staggering volume of information. An AI platform that can rapidly and reliably perform foundational tasks like measuring tumor volume and distance to the chest wall frees up the human expert to focus on higher-level synthesis, patient communication, and complex surgical judgment. According to some clinical experts, this human-AI partnership has the potential to not only improve efficiency but also enhance diagnostic confidence and reduce the risk of error.

The 3D visualizations also serve as a powerful communication tool. They can help multidisciplinary care teams—including surgeons, radiologists, and oncologists—arrive at a shared understanding of a case. Moreover, they can be used to explain complex surgical options to patients in a clear, visual way, empowering them to participate more fully in shared decision-making about their own treatment.

The Promise and Peril of Equitable AI

Beyond individual patient care, technology like this holds the potential to address broader issues of health equity. By encapsulating a high level of analytical expertise in software, AI could help democratize access to top-tier surgical planning, particularly in rural or underserved areas that may lack access to fellowship-trained breast radiologists.

However, the path to equitable AI is fraught with challenges. A primary concern is data bias. If AI models are trained on datasets that do not reflect the diversity of the global population in terms of ethnicity, age, and even breast density, their performance may suffer when used on underrepresented groups, potentially exacerbating existing health disparities.

Furthermore, the high cost of implementing and maintaining advanced technology can create a new digital divide, where only well-funded health systems can afford the latest tools. Without deliberate planning and thoughtful policy, the most advanced medical AI could become a luxury, widening the gap in care quality between the haves and have-nots.

Realizing the full, equitable potential of AI in medicine will require a concerted effort from developers, regulators, and healthcare systems. It demands the creation of diverse and representative training datasets, the development of affordable and accessible solutions, and the establishment of clear regulatory frameworks to ensure these powerful tools are deployed safely and fairly for the benefit of all patients.