Commentary: Land of the free, home of the hamburger

Born in 1969 and named after the lunar excursion module that landed that year in the Sea of Tranquility, Lem was destined for engineering. Knowing him as I have these last dozen years - a devoted sidekick to one of the maverick visionaries of medical imaging, my old friend Dimitri - Lem was clearly out of place among the crowd milling in front of the Seymour Community Museum.

"Do you have the time?" I asked, coming up beside him.

Lem, his mind elsewhere - as it so often is - and in a hurry to end the intrusion, looked straight at his watch.

"About 10 minutes . . . ," he said, before I cut him off.

"Until the museum opens," I finished, smiling.

Dimitri had sent him here. I had found this out from the staff at Dimitri's lab. They were willing to tell me how to find Lem, but swore they had no knowledge of Dimitri's whereabouts. I figured to find Dimitri through Lem, but first I had to find out why Lem had been sent to Seymour, WI.

I soon learned that Lem had made a mistake - a big one. He had not shown enough creativity to suit Dimitri. The little town of Seymour, home of the hamburger, was supposed to take care of that.

Here Lem was to seek out the roots of creativity, follow the culinary tendrils that led to an invention that has come to dominate American life and then, with a new understanding of this infrastructure of creative thought, he was to apply it - following Dimitri's lead, of course - to create the next great invention in medical imaging.

"What does Dimitri have in mind?" I asked.

"I haven't a clue," Lem said. "Obviously."

Lem's unworthy ideas had paved the way to Seymour, but his enthusiasm for them remained intact. He had proposed an idea called text mining to his mentor. Text mining allows computers to extract information from unstructured text.

Researchers at the University of California in Irvine developed a technique based on this idea to discern patterns of words and categorize them into topics. The researchers used "topic modeling" to analyze 330,000 stories, mostly published by The New York Times, to chart trends from apartment rentals in Brooklyn to voting irregularities. Lem thought the idea had merit outside the humdrum of everyday life.

"With text mining, we could analyze journal articles to identify trends in imaging," he said. "The first step would be to analyze archived articles, going back five or maybe 10 years in the National Library of Medicine's MedLine. We could compare them to actual procedure volumes, and see how long it took for different imaging techniques to catch on. If we could come up with equations that accurately chart what happened retrospectively, we could plug current trends data into those equations - and predict changes in practice patterns several years in the future."

Lem was breathless, but not nearly done. He made it clear this was just one idea.

Through half-sentences cut short by the weighty humidity that engulfed Seymour, Lem jumped into explaining AMIRA, a real-time decision support system developed by Paris-based Kaidara Software. Working through the AMIRA system, distributed-knowledge databases handle questions and respond with text-to-speech voice communications and laptop displays in text, HTML, and Flash graphics. Two demonstrations involving firefighting have led to impressive results.

"Imagine if we developed such a system for consultations on medical images. We'd ask questions and get voice replies supplemented by annotated images," he said. "Instant expert opinions!"

Lem suddenly fell silent. He was winded, but that wasn't the reason. His gaze ran along a crack in the pavement.

"Where I went wrong was in telling Dimitri why these ideas would work so well in medical imaging," he said.

Lem had observed that medical imaging trails behind other industries in adopting new technologies. So, he reasoned, to get ahead in imaging might require nothing more than coming abreast of the more advanced applications in other industries. Dimitri was appalled.

"I'd never seen him quite that angry," Lem said. "I think it's because he really . . ."

"Liked the ideas," I blurted. Lem gave me one of those looks. "Sorry. I shouldn't . . ."

"Finish other people's sentences?" Lem fired back.

The door to the museum swung wide. It was 1 p.m. on July 30, five days from the start of Seymour's BurgerFest, celebrating the great Charlie Nagreen. Charlie was just 15 years old in 1885 when he came up with the ground-beef-and-onion sensation that would mesmerize the U.S. more than a century later.

A 12-foot-tall likeness of "Hamburger Charlie" now stands guard over Seymour, one outstretched arm holding a hamburger to welcome visitors. Inside the museum was much more about Charlie and the town that fomented his creativity. But that was for Lem to discover, just as it would be for him to experience BurgerFest.

As for me, I needed to find Dimitri and learn what was driving him, and what had made him suddenly up and leave his laboratory.

While saying goodbye, Lem told me Dimitri had gone about 250 miles north to fish the lakes around Hayward, WI, where the world's largest muskelunge had been caught. If I was going to catch Dimitri, I would have to get going.