Machine learning, quantum computing can transform health care, including diagnosing pneumonia
The image on the left shows a nor­mal chest X‑ray, where­as the one on the right shows lungs with pneu­mo­nia opac­i­ty (Bre­viglieri, 2019). Cred­it: Fron­tiers in Com­put­er Sci­ence (2024). DOI: 10.3389/fcomp.2023.1286657

Pneu­mo­nia, an infec­tion in the lungs that caus­es dif­fi­cul­ty breath­ing, is most com­mon­ly diag­nosed through chest X‑rays. Typ­i­cal­ly, those chest X‑rays are read by radi­ol­o­gists, but work­force short­ages mean that in the future, it could be hard­er to get a diag­no­sis in a time­ly man­ner.

Addi­tion­al­ly, ear­ly and accu­rate diag­no­sis of pneu­mo­nia is impor­tant as it accounts for about 15% of deaths in chil­dren younger than 5 years old, accord­ing to the World Health Orga­ni­za­tion.

That’s where machine learn­ing comes in, said Srid­har Tayur, Ford Dis­tin­guished Research Chair and Uni­ver­si­ty Pro­fes­sor of Oper­a­tions Man­age­ment in Carnegie Mel­lon Uni­ver­si­ty’s Tep­per School of Busi­ness.

“Machine learn­ing is used for pre­dic­tion, and in health care we want to pre­dict if some­body has a dis­ease or not,” he said. “If you give enough exam­ples of images that have pneu­mo­nia and not pneu­mo­nia, because there are two cas­es, this is called bina­ry clas­si­fi­ca­tion.”

Tayur and a team of researchers stud­ied a tech­nique called sup­port vec­tor machine for clas­si­fi­ca­tion using quan­tum-inspired com­put­ing, then com­pared it to oth­er meth­ods in a paper appear­ing in Fron­tiers in Com­put­er Sci­ence.

“We showed that it is pret­ty com­pet­i­tive,” he said. “It makes few­er mis­takes and it takes less time.”

How can quantum computing be applied to health care?

Tayur found­ed the Quan­tum Tech­nolo­gies Group at CMU to bet­ter under­stand and apply quan­tum com­put­ing meth­ods to indus­tries such as health care.

“Peo­ple are always look­ing for more effi­cient ways of solv­ing prob­lems and nov­el meth­ods and tech­nolo­gies to tack­le it,” he said.

In the mid-20th cen­tu­ry, sci­en­tists who led the first quan­tum rev­o­lu­tion changed the world with inno­va­tions such as the tran­sis­tor, laser and atom­ic clock. While hard­ware to com­pute using qubits is still in devel­op­ment, sim­u­la­tors are capa­ble of tack­ling prob­lems of real­is­tic size with spe­cial­ly tai­lored algo­rithms, which is why this approach is known as quan­tum-inspired com­put­ing.


Cred­it: Carnegie Mel­lon Uni­ver­si­ty

“Assum­ing that qubit devices of larg­er size and low­er errors are going to be devel­oped, we can sim­u­late them on a reg­u­lar com­put­er right now,” Tayur said.

What are the challenges facing health care in adopting AI?

These tech­nolo­gies, how­ev­er, are still at the lead­ing edge of con­sid­er­a­tions when it comes to the appli­ca­tion of arti­fi­cial intel­li­gence in health care.

In order to do so, the indus­try has four chal­lenges ahead of it, as Tayur described in research with Tin­g­long Dai of Johns Hop­kins Carey Busi­ness School: physi­cian buy-in, patient accep­tance, provider invest­ment and pay­er sup­port.

To achieve these goals, any AI applied to health care sys­tems should con­sid­er how physi­cians will inte­grate it into their prac­tices, and then review how patients per­ceive the role of AI in health care deliv­ery.

“We wrote that paper in 2022, but things haven’t changed that much. It’s not just about build­ing a bet­ter mouse­trap, it’s about get­ting peo­ple to use that mouse­trap,” he said, ref­er­enc­ing a long-held busi­ness idea that suc­cess comes from sim­ply design­ing the best prod­uct.

First, as an exam­ple, Tayur explained that more than 500 med­ical AI devices have been approved by the FDA, but wide adop­tion of these tech­nolo­gies is still just begin­ning, in part because of the state of the health care indus­try and where finan­cial incen­tives lie.

“Hav­ing a good prod­uct is nec­es­sary, but it’s not suf­fi­cient,” he said. “You still need to fig­ure out how peo­ple are going to use it, and who is going to pay for it.”

Sec­ond, a major con­sid­er­a­tion in health care is lia­bil­i­ty. When it comes to devices, a com­pa­ny might encour­age doc­tors to adopt them, but what hap­pens if the device gives a faulty diag­no­sis or a doc­tor gives an incor­rect inter­pre­ta­tion of the data from the device?

“In the paper, we basi­cal­ly talk about the fact that you have to fig­ure out the busi­ness case, both risk and reward, along with train­ing and upfront invest­ments in adopt­ing the tech­nol­o­gy,” he said.

In apply­ing ele­ments of AI and quan­tum com­put­ing to health care, Tayur said while at least some progress has been made, there is still a long way to go.

“Many times what hap­pens is a lot of the AI in health care is derived by sci­en­tists and research physi­cians,” he said. “What they need is a busi­ness per­son who is less enam­ored by the mouse­trap and more sen­si­tive to the patient jour­ney and com­mer­cial via­bil­i­ty.”

More information:Sai Sakunthala Guddanti et al, Pneumonia detection by binary classification: classical, quantum, and hybrid approaches for support vector machine (SVM), Frontiers in Computer Science (2024). DOI: 10.3389/fcomp.2023.1286657Provided byCarnegie Mellon UniversityCitation:Machine learning, quantum computing may transform health care, including diagnosing pneumonia (2024, March 19)retrieved 3 April 2024from document is subject to copyright. Apart from any fair dealing for the purpose of private study or research, nopart may be reproduced without the written permission. The content is provided for information pur

poses only.