AI-powered device delivers “amazing physiologic information” to guide thrombectomy treatments

1241
Hal Rice

The Clotild smart guidewire system (Sensome), which uses artificial intelligence (AI) to provide real-time insights on clot composition, is showing early promise in the first-in-human CLOT OUT study, according to Hal Rice (Gold Coast University Hospital, Southport, Australia). Rice discussed this novel technology and the potential it offers in endovascular thrombectomy procedures using electrical impedance spectroscopy at LINNC Paris 2022 (30 May–1 June, Paris, France).

The presenter started by noting that the first-pass effect is “very important” in this space, increasing the likelihood of a thrombolysis in cerebral infarction (TICI) 3 score and a “great clinical outcome” for the patient. Rice noted that recent studies have also confirmed that clot composition has a significant bearing on patient outcomes too. Red blood cell-rich thrombi, for example, are more susceptible to distal embolisation, while fibrin-rich and platelet-rich thrombi may be harder to remove but also have less chance of embolising distally.

Moving on to discuss the Clotild technology, Rice noted its unique, “very clever” microsensor, which is 2mm in size and performs impedance measurements that are transmitted to the Cloviz tablet device via a Bluetooth transmitter, providing real-time insight on what the clot looks like, and the length of the occlusion. This translates to the sensor delivering higher impedance measurements in red blood cell-rich clots, and lower impedance measurements for fibrin-rich clots, he added, reporting a “really good differentiation” between the two.

Rice stated that, owing to the nine sensors on the device, arranged into three rows, it is also capable of identifying clots that have a ‘mixed’ composition—which may have substantial amounts of red blood cells, fibrin and platelets in their makeup—with this being signified by different impedance readings between the different groups of sensors.

He then detailed that the Clotild device uses a database centred around clots with a known composition. This database has been used to train a machine-learning model capable of recognising different compositions, which is then able to predict composition in unknown clots, Rice reported. He also claimed that this use of electrical impedance spectroscopy presents a novel means for characterising occlusions in stroke, with data having shown continued coherence through ex vivo, animal and early in-human research thus far.

The next part of Rice’s presentation homed in on the CLOT OUT trial, which is the first in-human study evaluating the safety of the Clotild smart guidewire system. Training and validating the aforementioned prediction model via a database is also a key objective of CLOT OUT, he noted. The single-arm, multicentre study began in late 2021, is taking place across Australia and Europe, and intends to enrol roughly 100 patients with an M1-origin acute ischaemic stroke in order to collect a total of 60 clots. Providing a status update on CLOT OUT, Rice said the trial has enrolled 11 patients to date, with no device-related complications being observed in these completed cases, and the first Data Safety Monitoring Board meeting recently recommended continuation of the study.

Concluding his talk, Rice said this “exciting” technology is akin to other stroke care innovations like robotics and administering thrombolytics post-thrombectomy. “As well as what we see on an angiogram or feel in terms of resistance during a procedure, this wire means we are now getting amazing physiologic information too,” he continued. “It also integrates very well into our current mechanical thrombectomy workflows, with very little delay to these time-critical procedures.” Here, he reported that deployment of the Clotild device adds fewer than five minutes to the overall thrombectomy workflow.

Rice also alluded briefly to future applications beyond clot composition and length, such as in vessel-wall analysis to elucidate other key occlusion characterises like smooth muscle content and cholesterol—possibly identifying clots that are actually intracranial atherosclerosis (ICAS) lesions, not thromboembolic occlusions.

“These parameters can really help to guide the mechanical thrombectomy procedure and potentially increase the first-pass effect going forward,” he stated. “It is early days, with 11 patients out of 100 enrolled, but the technology is definitely very promising.” The Clotild device and its potential were well-received at LINNC, with Jacques Moret (Bicêtre University Hospital, Paris, France) noting that it is a “fascinating” use of high technology that may help to answer questions regarding what happens distally to a clot following its removal. This point was echoed by Vitor Pereira (St Michael’s Hospital, Toronto, Canada), who added that the device has the potential to improve first-pass recanalisation rates by “bringing personalised medicine to the thrombectomy world”.

Rice presenting at LINNC Paris 2022

LEAVE A REPLY

Please enter your comment!
Please enter your name here