Qure.ai's qER: The Globally Proven Algorithm for Head CTs

AI biases can emerge if algorithms are not validated on diverse populations.

qER, Qure.ai's AI solution for head CTs, has been rigorously evaluated using real-world, multi-regional data. As a result, qER reliably delivers excellent outcomes as a fully operational product for diverse populations globally. It has been extensively tested, deployed, and proven accurate across diverse patient demographics and scanning equipment, instilling true confidence in qER's reliability anywhere.

Qure's FDA-cleared qER software has gained traction in North America for its ability to swiftly and accurately triage, detect, and quantify pathology on head CTs. Covering four critical markers - hemorrhage, mass effect, midline shift, and fractures - qER delivers comprehensive insights to prioritize critical cases.

Monitoring severity over time aids in the long-term management of conditions such as hemorrhagic stroke and hydrocephalus. Input from North American hospitals and teleradiology groups guided qER's emphasis on essential CT measurements for emergency triage.

Additionally, qER-Quant leverages intelligent algorithms to precisely quantify volumes of intracranial structures and lesions. By determining the severity and progression of conditions like hemorrhagic stroke and hydrocephalus, qER-Quant further augments diagnostic and treatment decisions.

A study reviewed by vRad validated the qER algorithm against radiologist findings on more than 300,000 head CT scans. In one of the largest* evaluations conducted, qER delivered an outstanding AUC of 0.96, showcasing its ability to reduce door-to-scan interpretation times by up to 96% and accelerate communication of critical results to stroke teams within 5 minutes.

*Based on data size

A noteworthy study published in The Lancet demonstrates a crucial validation of deep learning's capabilities in detecting urgent abnormalities on CT scans, using a diverse 300k scan dataset sourced from 20 centers in India. Key findings demonstrate exceptional diagnostic accuracy, with AUC values of 0.92 for detecting intracranial hemorrhage and over 0.90 for other specific bleed types.

Aarthi Scans and Labs, one of India's largest diagnostic chains, has adopted AI to enhance their teleradiology reporting - including for remote Tier 3 cities lacking nighttime radiologist coverage. This enables faster diagnosis and treatment for time-sensitive trauma cases. One deployment at Aarthi's Tenkasi center showed great insight into AI's clinical impact.

In Tenkasi, patients with conditions like stroke or traumatic brain injury used to wait over 7 hours

from the onset of symptoms to receive treatment due to delays in scan analysis, transfer to tertiary hospitals, etc.

According to Dr. Arunkumar Govindarajan, Aarthi's Director, "qER reduced symptom-to-treatment time by 30-70%, allowing more patients to be managed locally. By getting AI's assistance for faster diagnosis 24/7, cities like this could revolutionize outcomes for time-sensitive neurological injuries and disease."

A study detailed in Diagnostic Neuroradiology further cements the value of qER, focusing on the crucial aspect of intracerebral hemorrhage (ICH) management—hematoma volume quantification. Hematoma volume is a pivotal predictor of patient outcomes post-ICH, and the study aimed to validate a novel, fully automated software developed by Qure.ai for this purpose.

The study analyzed a significant patient population from the Swedish Stroke Register, including all patients diagnosed with intracerebral hemorrhage (ICH) in Region Skåne from 2016 to 2019. It compared ICH volume measurements from qER–NCCT, Qure.ai's automated quantification tool, against traditional methods like ABC/2 and manual segmentation using Sectra's software.

The results demonstrate qER–NCCT's ability to identify ICH with 97% sensitivity, proving highly efficacious in detecting bleeding. Volume measurements also showed excellent concordance between qER–NCCT and manual segmentation, considered the gold standard, with an interclass correlation of 0.96.

qER has gained widespread recognition and adoption within the United Kingdom's National Health Service (NHS) for its ability to expedite the identification of critical markers on head CT scans. Implemented in over 20 NHS trusts and health boards spanning England, Scotland, and Wales, qER processes over 1 million emergency scans every year, demonstrating high accuracy and reliability.

qER achieves exceptional AUC scores above 0.95 for the detection of intracranial hemorrhage and other hemorrhages. The deep learning algorithms also showcase near-perfect correlation with specialized neuroradiologist reporting for hematoma volume quantification, with an interclass correlation coefficient of 0.96.

Moreover, Medica, a leading teleradiology provider in the UK, has integrated Qure.ai's qER solution into its acute imaging workflow. qER serves as an "augmented intelligence" tool that analyzes non-contrast CT scans of the brain, flagging potential abnormalities and allowing for priority reporting based on clinical urgency.

As Dr. Garry Pettet, Consultant Radiologist, explains: "The qER algorithm is handy to reporting clinicians when interpreting acute brain CTs. It not only facilitates the rapid allocation of critically abnormal scans to our reporters but also helps to increase our reports' accuracy by lowering human error."

Recognizing the immense potential to augment emergency care pathways, the NHS selected qER as one of the inaugural winners of the prestigious Artificial Intelligence in Health and Care Award. The £140 million award spans 4 years to assess promising AI innovations in various clinical settings thoroughly. For qER, the aim is to widely implement it in Accident and Emergency (A&E) departments to enhance diagnostic turnaround times for critical conditions like hemorrhagic strokes, ischemic strokes, and traumatic brain injuries.

The extensive piloting of qER technology within the NHS reflects an understanding of the pivotal role AI can play in bolstering overburdened care teams. With A&E departments experiencing surging footfalls and exponential growth in required imaging, advanced analytics offered by solutions like qER could help alleviate clinical workload while boosting diagnostic precision.

Integrating AI-driven measurements into clinical practice could revolutionize diagnosing and monitoring neurologic and craniofacial conditions. It would provide a standardized reference supporting more informed decision-making and follow-up. Furthermore, initiatives like the AI-Based Stroke Care Network launched by Zydus Hospitals and Medtronic in Gujarat, India, aim to bridge significant gaps in stroke care, particularly in regions with limited access to specialized healthcare facilities, by leveraging AI to interpret CT scans remotely. Such innovations are pivotal in enhancing the timeliness and efficacy of stroke treatment, thereby improving survivability and reducing the incidence of long-term disability. 

qER's seamless integration into healthcare systems across continents has accelerated the detection of neurologic emergencies, notably reducing response times and facilitating timely interventions. Through ongoing innovation and partnerships, algorithms like qER stand poised to radically improve healthcare delivery, offering tailored, efficient, and equitable medical solutions on a global scale.