Super computer, Crowd-sourcing play key roles in the investigation at the 26th annual Historical Clinicopathological Conference.

In a battle pitting man versus machine, physicians around the world and a supercomputer rendered their diagnoses on what caused the death of a saint who lived 800 years ago.

For the first time in its 26-year history, the Historical Clinicopathological Conference (CPC) used the popular crowd-sourcing technique to poll more than 500 physicians in 46 countries on how St. Francis of Assisi died and compared their top answer to the intelligence of a supercomputer.

“The CPC has always been a fascinating exercise probing what caused some of the most celebrated figures in history to die,” said Dr. Philip Mackowiak, the CPC’s founder and emeritus professor of medicine and the Carolyn Frenkil and Selvin Passen History of Medicine Scholar at the University of Maryland School of Medicine. “But pitting man against machine adds another interesting element to the conference and takes it to a new level of investigation.”

St. Francis developed fevers, chills, severe abdominal pain, delirium, blindness, open sores and other symptoms over a period of more than 20 years while living in Europe and traveling to north Africa. What could have caused such suffering?

Eliot Siegel, MD, professor and vice-chair of diagnostic radiology and nuclear medicine at the University of Maryland School of Medicine, and an early supporter of the use of super computing in medicine, and medical crowd-sourcing expert Reza Manesh, MD, a hospitalist and assistant program director for clinical reasoning for the Osler Medical Training Program at Johns Hopkins University School of Medicine, presented their diagnoses at the 26th annual Historical Clinicopathological Conference on Friday, May 3, at the University of Maryland School of Medicine’s Davidge Hall.

Dr. Siegel fed information into the Isabel DDx, a super computer used by physicians around the world to help improve their accuracy when diagnosing diseases. Isabel received only the information Dr. Siegel submitted into the system as if the patient were alive today. The specifics included, age, 40 to 49 years old. Travel history, Egypt. Illnesses, paresthesia, which can manifest as a numbness of the skin including the hands, feet and face; skin lesions, eye discharge, photosensitivity. Within seconds he received his answer: leprosy, most likely borderline tuberculoid.

“What surprised me was Isabel came up with the diagnosis from those symptoms without knowing the historical background, especially without knowing that the historical figure had a ‘special affinity for lepers’ and actually washed the skin and the wounds of lepers,” said Dr. Siegel, who noted that Isabel spit out an answer in a second.

Dr. Manesh entered St. Francis’ information into the Human Diagnosis Project, an open collective intelligence system, to seek answers. Its contributors are more than 20,000 medical professionals-- primarily doctors, residents and students across a broad spectrum of specialties. Each contributor provides insight independently and can’t see suggestions from other users.

“By combining insights from multiple physicians with diverse perspectives … into a collectively weighted list, you’re more likely to arrive at the right decision,” Dr. Manesh said.

Dr. Manesh’s crowd-sourcing supplied the same answer: leprosy. Before he submitted the case to the crowd, Dr. Manesh originally thought it was a different illness, Behcet’s. But the responses prompted him to do more reading, and he changed his final diagnosis.

“I thought it would be Behcet’s, but I believe in the collective wisdom of the group and therefore I changed my final answer to leprosy. This was only after seeing the group’s response. I’m a firm believer in the crowd,” Dr. Manesh said.

Sister Joanne Schatzlein, a member of the Sisters of Saint Francis of Assisi, who also presented at the CPC, spent nearly eight years in the 1980s – three working with Daniel P. Sulmasy, MD - researching whether St. Francis had leprosy. When she raised the question, it was the first time the leprosy theory was presented, she said. She and Dr. Sulmasy published their findings in St. Bonaventure University’s Franciscan Studies. The book was republished in 2014 under the title "Francis the Leper, Faith, Medicine, Theology, and Science."

“I’m just so impressed our research has been supported in a pretty high-tech level,” Sr. Schatzlein said. "I just laughed out loud, thinking how long we researched and pondered, and the computer just spit it out. Dan and I will have to celebrate again!"

Monthly, physicians and medical students at the School of Medicine review modern unusual medical cases to learn how experienced clinicians approach cases that are difficult to diagnose. Such reviews are known as clinicopathological conferences.

Once a year, however, the University of Maryland School of Medicine chooses an historical figure to puzzle out the pieces of their mysterious illness. Past years included examining the symptoms and medical histories of Alexander the Great, Herod, Charles Darwin, Florence Nightingale, Oliver Cromwell, and others, to develop diagnoses or discuss end of life issues.

Conference organizers used the 800-year-old case to demonstrate how in the future world of medicine, man and machine will work together to treat ill and injured patients.

“The CPC offers the potential to reach a fairly wide audience. It will help train the medical students of the future. We are constantly exploring how to keep the medical school curriculum as relevant as possible not just today, but to where we will be in 10 to 20 years,” said Dr. Siegel, who has worked with IBM to apply supercomputer Watson’s technology to medicine, and is chief of imaging for Veterans Affairs Maryland Healthcare System and VISN 5.

“It is a really fascinating and timely challenge between the supercomputer on the one side … and on the other hand the consensus of experts reflecting the trend toward self-driving cars, speech recognition, personalized assistants, and other forms of artificial intelligence,” Dr. Siegel said.

While crowd-sourcing is best known for raising funds for charities or new businesses, a version using collective intelligence and machine learning also has taken root in the medical field. It can pull together the knowledge and experience of a large number of health care providers to reach a more enlightened answer in a particular case. While about 85 percent of the time physicians can correctly diagnose a patient on their own, a not insignificant minority of cases can be elusive. That’s where collective intelligence can be most helpful, said Dr. Manesh, who is also an assistant professor at the Johns Hopkins University School of Medicine.

Dr. Siegel sees a future where computers will interpret radiology images and make recommendations on patient therapies based on lab findings and genomic data.

But both Dr. Siegel and Dr. Manesh dismiss the notion that machines will replace physicians.

“The computer will never replace the physician, I think in my lifetime, but it will be a good supplement,” Dr. Manesh said. “I like artificial intelligence, but it’s not going to put me out of business. Human Dx is unique in its open, inclusive approach to AI because it has been explicitly designed by and for medical community members like myself to augment our capabilities.”

Dr. Siegel agreed. Instead of computers replacing physicians, they will be used to enhance care that physicians provide.

“I do not believe any of us, including today’s medical students, will be in any way replaced by computers,” Dr. Siegel said.

“The future of medicine is man and machine working together and there are so many possibilities that we have to explore,” Dr. Siegel said. “It is a partnership that will undoubtedly shape the future of medicine.”

About the CPC

“The Historical CPC” was developed in 1995 by Philip Mackowiak, ’70 emeritus professor of medicine and the Carolyn Frenkil and Selvin Passen History of Medicine Scholar. Since then, the CPC has been administered by Dr. Mackowiak and produced annually by the University of Maryland School of Medicine, Baltimore VA Medical Center and the Medical Alumni Association. A CPC, or Clinicopathologic Conference, is a medical clinical exercise in which the history of a patient’s illness is presented to an experienced clinician for discussion in a didactic setting. This form of conference is used to teach students and house staff how an experienced clinician would approach a difficult or challenging case. We present an unusual, modern case on a weekly basis, but once a year we stray from our modern cases and discuss an historical figure.