A new imaging technique that pairs acoustic resolution photoacoustic microscopy co-registered with ultrasound and an artificial intelligence neural network was proven feasible for use in detecting residual tumor tissue in patients with rectal cancer.
An imaging technique capable of differentiating between residual cancerous versus noncancerous rectal tissue following radiation and chemotherapy has been validated by a team of investigators whose research was recently published in Radiology.1
These methods, comprised of an endorectal co-registered photoacoustic microscopy (PAM) and ultrasound (US) system coupled with a convolution neural network (CNN), may offer the possibility of avoiding unnecessary surgery in patients achieving complete tumor destruction after initial therapy.
“Our PAM/US system paired with the deep learning neural network has great potential to better identify patients suitable for nonoperative management and improve patient quality of life,” study author Quing Zhu, PhD, professor of biomedical engineering at the McKelvey School of Engineering, said in a press release.2 “If we can tell after radiation and chemotherapy which patients may have a good response with no residual tumors, the patient may be able to avoid surgery.”
In the prospective clinical trial (NCT04339374), patients completing radiation from September 2019 through September 2020 were included with images obtained with the PAM/US system before surgery. The PAM CNN and US CNN systems were trained to distinguish malignant from normal colorectal tissues by ex vivo (n = 22) and in vivo (n = 5) patient samples, with an additional 5 patients included for fine-tuning the neural network.
Imaging markers unique to PAM indicating complete tumor response were identified, which included the recovery of normal submucosal vascular architecture within the treated tumor bed. The model was able to capture this process and properly differentiate tissue of the residual tumor. An area under the receiver operating characteristic curve of 0.98 (95% CI, 0.98-0.99) was achieved in the 5 patients with PAM CNN versus 0.71 (95% CI, 95% CI, 0.70-0.73) with US CNN, which misclassified 3 out of 5 patients.
“This is spectacular news, and it moves us closer in the transition from concept to clinically useful technology,” Matthew Mutch, MD, the Solon and Bettie Gershman Chair in Colon and Rectal Surgery, chief of the Section of Colon and Rectal Surgery and professor of surgery at Barnes-Jewish Hospital in St. Louis by School of Medicine, said in a press release. “The hope is that it will allow us to differentiate patients who had a complete response to chemotherapy and radiation from those patients with residual tumor. This will help better determine which patients can be managed nonoperatively versus those who truly need an operation.”
The team developing the technology say they spent over 3 years investigating this technology in surgically removed colon and rectum specimens prior to developing the prototype for patient studies.
Patients under anesthesia are imaged using a handheld endorectal laser probe with a rotating head that takes a 360-degree image of the rectum as well as the last 6 inches of the colon. One image per second is taken from the end of the probe which is engulfed by a small latex balloon filled with water that allows for transmission of ultrasound and photoacoustic waves to the rectal wall, which highlight changes to the vasculature in new tumor growth and other tissues. This technology may serve an unmet medical need, since MRI imaging is often unable to determine new or residual tumor tissue for scar tissue in the treated area.
“From the very preliminary ex vivo data, my setup clearly disclosed multi-layer structure from ultrasound image and rich blood vessels in the submucosa of normal colorectal tissue,” Leng said. “In contrast to normal tissue, malignant tumor bed shows a lack of multilayer structure and blood vessels. This important finding may reveal an important feature accessing patients” treatment response to chemotherapy and radiation therapy.”
1. Leng X, Uddin KMS, Chapman W Jr, et al. Assessing rectal cancer treatment response using coregistered endorectal photoacoustic and US imaging paired with deep learning. Radiology. March 23, 2021. doi: 10.1148/radiol.2021202208
2. Leap forward’ in risk management of rectal cancer. News release. Washington University in St. Louis. March 25, 2021. Accessed April 1, 2021. https://bit.ly/3cGet0E