People who escape fatal fires with their lives may face a tangle of emotions: fear, regret, relief.
In many cases, they also face criminal charges.
If just 5 percent of the nation's half-a-million yearly structure fires are suspicious, Lentini said, then that means 25,000 chances to mistakenly charge someone with arson.
"And if they [prove a crime took place] with bad evidence, and the jury believes that it's a set fire, many times there is no doubt about who did it," he said.
Kristine Bunch was convicted by a jury in 1996 of setting a fire in her Indiana trailer that killed her 3-year-old son. She received concurrent sentences of 50 years for arson and 60 years for murder.
"This is a woman who has no prior criminal history ... no arrest record ... no psychiatric history," said Jane Raley, senior staff attorney at the Center on Wrongful Convictions at Northwestern University School of Law. "There was nothing here."
Raley is working to get Bunch, 36, a new trial. "There is no motive," Raley said. "Why would someone like Kristine do this?"
Lentini questioned the evidence against Bunch.
"I can turn just about any fire into an arson fire if that's what I want to do," he said.
We brought Lentini to the Fire Science Program at Eastern Kentucky University, where technicians conduct fire tests in a bunker set up with rooms of ordinary furniture.
Fires are set in the bunker not with gasoline or other flammable liquids but with just a spark, as if from a burning cigarette.
"We are hoping to generate some of the artifacts that people in the past have called evidence of arson," Lentini said.
Among those whom Lentini counts as having believed misguided arson evidence in the past is ... John Lentini.
Twenty years ago, Lentini got a life-changing lesson in arson fires. Back then, he was convinced that a man named Gerald Lewis had killed his pregnant wife and four children by setting a fire with gasoline in Jacksonville, Fla.
"I was scheduled to be deposed the next morning for the prosecution," he recalled. "I was going to help them send Gerald Lewis to Old Sparky [the electric chair]."
But before he testified, Lentini had a chance to test Lewis' claim that the blaze began by accident when his couch caught on fire.
Lentini's test is now commonly called the "Lime Street Fire Test." To Lentini's surprise, the fire in the test left behind the exact same signs as if it had been started with gasoline, even though none was used.
"It was an awakening," Lentini said.
Which is why Lentini is now driven to bring hard science into what traditionally has been seen as the "art" of fire investigation.
"I think when they said ... 'art,' I think what they meant was luck," Lentini said.
Back at Eastern Kentucky University, investigators began their fire experiments, with cameras and thermal imaging sensors recording every moment.
The tiny couch fire grew steadily as a layer of hot dense smoke formed on the ceiling. The sensors showed how hot it was. Temperatures reached about 1,200 degrees. Soon, blazing hot gas started to descend.
Just minutes after the fire was started, the hot gases suddenly ignited, and the room exploded.
This is a phenomenon known as "flashover."