Tencent Medical"Tencent Medical" is an artificial intelligence medical imaging platform launched by Tencent, which uses artificial intelligence technologies such as image recognition, d...
At 3 a.m. in the imaging department of a top-tier tertiary hospital, Dr. Li, on night duty, rubbed his tired eyes while staring at the computer screen. Nearly 300 CT scans were displayed on the monitor, each containing details that required careful examination—even millimeter-sized nodules had to be spotted among the dense lung textures. He had just found the third suspected early-stage lung cancer lesion, glanced at the wall clock, and knew he would likely stay up until dawn that night; even three cups of strong coffee couldn’t suppress his fatigue.
Just then, he clicked on a system tool labeled “Tencent Medical”. This was something he had encountered at the department training meeting last week. He didn’t have high expectations at first, but unexpectedly, after uploading the last batch of scans, in less than three minutes, the system popped up three suspicious areas marked with colours. One of them overlapped with the small nodule he had just found, and the other two were tiny ground-glass opacities he hadn’t noticed earlier. Dr. Li stared at the screen for a few seconds, paused his finger on the mouse, and finally clicked confirm on the system’s marked points—this would save at least an hour of reading films that night.
Official website: https://medical.qq.com/medical
In fact, for Dr. Li, he doesn’t need to understand the algorithm logic of this AI; he only needs to know how to import the scans he has. His daily workflow hasn’t changed much: arriving at the department at 8 a.m., first organizing the imaging data to be read that day, classifying CT, fundus or endoscopy images by patient number, then batch importing them to the backend of Tencent Medical through the system’s upload port. The whole process isn’t much different from other medical assistance tools he used before, except for the waiting time—previously, batch analysis took more than ten minutes, but now results are available in just three to five minutes.
The results given by the system don’t directly generate a diagnosis report; instead, they mark suspicious areas with boxes of different colours, accompanied by preliminary risk ratings and reference opinions. Dr. Li will first compare the positions marked by the system with his own reading results, and when encountering uncertain areas, adjust his judgment based on the system’s suggestions. He has tried several times that the tiny lesions marked by the system are almost the ones he almost missed due to time pressure, which gives him more confidence when writing diagnosis opinions for patients.
The department where Dr. Li works counted a set of data last month: previously, on average, each doctor spent three and a half hours a day reading films; now after using Tencent Medical, this time has been compressed to about two hours, and the remaining time can be used to talk to patients more about their conditions, or organize previous cases. He said the most obvious change is that he no longer has to focus all his energy on finding lesions, and can pay more attention to communicating with patients. After all, when patients come to the hospital, what they need most is the doctor’s patient explanation, not the doctor’s back facing the screen.
But he also said that this AI is not meant to replace doctors, but to help doctors share those repetitive and error-prone tasks. For example, some early tiny lesions—even experienced doctors may suffer from visual fatigue after long hours of reading films and miss one or two, and the system can help mark these easily overlooked areas, equivalent to giving doctors a pair of “non-tired eyes”. He also tried to compare the system’s results with the diagnosis of senior colleagues; in most cases, the suspicious areas marked by the system were confirmed by colleagues, which made him trust this tool more and more.
When Tencent Medical was just launched, Dr. Li also read relevant introductions online, knowing that this is Tencent’s first product applying AI to the medical field, with the core goal of helping doctors improve diagnostic efficiency and accuracy. He still remembers a sentence that impressed him when he first listened to the product manager’s explanation: “We made this not to let machines replace doctors, but to free doctors from repetitive physical work, so that they can spend more time on places that need humanistic care.”
He later gradually found that this system is highly adaptable; whether it’s a primary hospital or a tertiary hospital, it can adjust the modules used according to its own needs. For example, doctors in some primary hospitals with less experience can refer more to the system’s suggestions; for large hospitals like theirs, it’s more used to reduce missed diagnoses and save time. He also heard that this system can not only analyse CT scans, but also fundus and endoscopy images, and even assist in the diagnosis of some common cancers. However, the department where he works currently mainly uses it for reading lung CT scans, and other functions haven’t been deeply explored yet.
Dr. Li is now used to opening Tencent Medical every day, importing the day’s imaging data, and watching the system mark those easily overlooked lesions. He said this isn’t some “black technology”, but something that really saves doctors a lot of trouble, allowing him not to stay up so late and see patients more at ease.
