Korean Intellectual Property Office (KIPO) registersinformation-providing system for cancer patient admission decision andtreatment goals
[Press9] A technology that predicts the riskof death of cancer patients based on medical images and medical voicerecordings has been registered as a patent.
The Korean Intellectual Property Officerecently approved Oncosoft's application for a patent titled 'Method forproviding information on mortality risk and device using the same'.
Oncosoft is a medical device company thatdevelops AI-based cancer treatment software (SW). In 2021, the company receiveda pre-series A investment from medical AI public company Buno.
The registered patent is a technology thatutilizes medical images, image readings, and medical voices to measure the riskof death for cancer patients. The research was supported by the Ministry ofScience and ICT and the Korea Research Foundation.
In the patent description, ONCOSOFT said,"Cancer patients need continuous observation of their condition. Cancertreatment that combines various treatment methods contains many side effectsand is more likely to recur than other diseases, so prognosis monitoring thatconsiders multiple risk factors is necessary.
In particular, the decision to admit acancer patient to a hospital for treatment, and the decision to predict thepotential for life-sustaining treatments, is a difficult and critical one.
"Accurate prognosis is crucial toensure that patients with a good chance of recovery are given the opportunityfor intensive care in the ICU, and to alleviate the suffering of patients andtheir families who are unlikely to recover," said Oncosoft.
Therefore, there is a need for a systemthat can provide information on prognosis and risk of death before settingtreatment goals and making decisions.
"We have noted that changes in medicalimaging readout data in EMR data can be associated with poor outcomes, such asdeath and poor recovery," says Oncosoft, "Consequently, medicalimaging readout data can be used to predict the risk of death for patients,particularly cancer patients."
They also found that data from patients'outpatient conversations could be used to predict mortality risk. We believethat medical transcripts should be more actively considered for mortality riskprediction.
To that end, the patent describes thecreation of a prediction model that uses medical imaging data and medical voicerecordings as inputs and outputs mortality risk as outputs.
"(The system) can provide informationon whether to admit cancer patients to the hospital, recognize emergencies inadvance by quickly identifying the patient's condition, and improve treatmentperformance by accelerating the time of treatment," said Oncosoft."It will contribute to reducing the cost of unnecessary treatment andincreasing the survival rate of cancer patients."