FDA Clears Second TULSA-AI Module Developed by Profound Medical

Profound Medical Corp. has received U.S. Food and Drug Administration (FDA) 510(k) clearance for its second TULSA-AI module, Contouring Assistant, to be used in conjunction with the TULSA-PRO system.

The Transurethral Ultrasound Ablation (TULSA) procedure could potentially become a mainstream treatment modality across the entire prostate disease spectrum; ranging from low-, intermediate-, or high-risk prostate cancer; to hybrid patients suffering from both prostate cancer and benign prostatic hyperplasia (BPH); to men with BPH only; and also, to patients requiring salvage therapy for radio-recurrent localized prostate cancer. TULSA employs real-time MR guidance for pixel-by-pixel precision to preserve prostate disease patients’ urinary continence and sexual function while killing the targeted prostate tissue via a precise sound absorption technology that gently heats it to kill temperature (55°C to 57°C). TULSA is an incision- and radiation-free “one-and-done” procedure performed in a single session that takes a few hours. Virtually all prostate shapes and sizes can be safely, effectively, and efficiently treated with TULSA, according to Profound Medical. There is no bleeding associated with the procedure; no hospital stay is required; and most TULSA patients report quick recovery to their normal routine.

“The addition of a second, FDA-cleared TULSA-AI module is an important milestone as we continue to deliver on our promise to continuously improve the TULSA treatment experience for urologists and their patients,” Profound Medical Chairman/CEO Arun Menawat, Ph.D., stated. “To that end, we believe that Contouring Assistant should not only increase urologists’ confidence in their treatment designs, but also significantly increase their procedural efficiency.”

Contouring Assistant, Profound’s second FDA-cleared TULSA-AI module, is a machine learning-based prostate segmentation tool that helps delineate the prostate and target ablation volume. This module uses a large database of physician-created TULSA treatment designs and recommends a design in a new procedure based upon that knowledge. Moving forward, as the TULSA treatment and outcomes database continues to grow through real-world adoption, and as artificial intelligence (AI) technology evolves, the proposed treatment designs are expected to continue to improve.

In support of the FDA 510(k) application, Contouring Assistant was subjected to several rigorous clinical analyses. First, a Reference Standard was created by combining contours from three expert prostate radiologists: Edward Steiner, M.D. (WellSpan York Hospital); Steven Raman, M.D. (UCLA Health); and Robert Princenthal, M.D. (Rolling Oaks Radiology/RadNet). Then, Contouring Assistant was validated in two clinical studies:

1. A standalone study in 100 prostate cancer cases, comparing Contouring Assistant to the Reference Standard. This study demonstrated that Contouring Assistant was non-inferior to the Reference Standard, similar to expert radiologist inter-reader variability; and contoured the 100 prostate cases with an average Sørensen–Dice coefficient of 0.910 +/- 0.031.

2. A multi-reader, multi-case (MRMC) study conducted by five radiologists—Pejman Ghanouni, M.D., Ph.D. (Stanford Medicine); Aytekin Oto, M.D. (University of Chicago); Adnaan Moin, M.D. (WellSpan York Hospital), Sandeep Arora, MBBS (Yale School of Medicine); and Ara Karamanian, M.D. (HALO Precision Diagnostics). The same study design was used to assess performance by three urologists: Laurence Klotz, M.D. (Sunnybrook Health Sciences Centre); Christian Pavlovich, M.D. (Johns Hopkins University); and Preston Sprenkle, MD (Yale School of Medicine). This MRMC study demonstrated that readers using Contouring Assistant were non-inferior in terms of contouring accuracy, but also 29% faster on average; and the three urologists who tested this feature showed statistically significant improvement in their segmentation accuracy using Contouring Assistant, and were 32% faster on average.

“Profound is to be commended for supporting this rigorous clinical analysis of the Contouring Assistant feature, which was on par with that normally associated with testing diagnostic-level AI software,” Sprenkle said. “Importantly, Contour Assistant not only allowed my colleagues and I to approach the accuracy of an expert radiologist reader in our TULSA treatment designs, but also enabled us to reduce overall procedure times by one-third.”

Profound is continuing development work on the next planned TULSA-AI module, TULSA BPH, and expects to provide more details on that later this year.

Profound Medical is a commercial-stage medical device company that develops and markets customizable, incision-free therapies for diseased tissue ablation. Profound is commercializing TULSA-PRO, a technology that combines real-time MRI, robotically-driven transurethral ultrasound and closed-loop temperature feedback control. TULSA-PRO is designed to provide customizable and predictable radiation-free ablation of a surgeon-defined prostate volume while actively protecting the urethra and rectum to help preserve the patient’s natural functional abilities. TULSA-PRO is CE marked, Health Canada approved, and cleared by the U.S. FDA.

Profound is also commercializing Sonalleve, a therapeutic platform that is CE marked for treating uterine fibroids and palliative pain treatment of bone metastases. Sonalleve has also been approved by the China National Medical Products Administration for the non-invasive treatment of uterine fibroids and has FDA approval under a Humanitarian Device Exemption for the treatment of osteoid osteoma. The company is in the early stages of exploring additional potential treatment markets for Sonalleve where the technology has been shown to have clinical application, such as non-invasive ablation of abdominal cancers and hyperthermia for cancer therapy.