Two siblings from Mililani High School have written a computer program to identify cancer in tissue samples, potentially aiding pathologists in the early detection of the disease. The effort earned the pair top team honors in the regional finals of a prestigious national science competition.

Sophomore Lucia and junior Philip Mocz will share a $6,000 scholarship and advance to the national Siemens Competition in Math, Science and Technology next month, competing against five other teams for scholarships ranging from $10,000 to $100,000.

The intent of their project was to find an automatic approach to cancer diagnosis, which is now a time-consuming and costly process that requires a human expert.

Their software program essentially makes cancer cells more visible in photographs of tissue samples.

By measuring the complexity and randomness of tissue samples, the program the two wrote is as accurate as other methods of cancer diagnosis currently used by pathologists.

Their 71 percent success rate in identifying lung cancer is competitive with molecular clinical diagnosis methods, said Andre Bachmann, an assistant professor at the Cancer Research Center and their project mentor.

"They developed a new set of tools for future research and diagnostics," Bachmann said. "It won't replace (current methods), but it will be an alternative approach."

The project — Computer-Aided Identification of Cancer from Photomicrographs by Entropy Analysis — evolved from sophomore Lucia's freshman science project, which won second place in the state science fair last spring.

"Then it was really in its early stages," she said. "Over the summer, through some ideas by Philip, we decided to develop into something more 'wow!' "

Philip, who beat Lucia for the top spot at the state fair, helped push Lucia's project into national contention at the regionals at Stanford University over the weekend.

Their computer program allows for the comparison of tissue samples to predict whether there is cancer present.

"The idea behind it is that you try to basically distinguish normal tissue from cancer tissue," said Bachmann, who provided the tissue samples.

Bachmann noted that he's a cancer scientist, not a mathematician, so he was not involved in developing the algorithms and calculations. He helped Lucia by giving her tissue samples, and she and Philip used Bachmann's suggestions to help draw their conclusions.

Lucia sees the program becoming part of regular doctor's checkups.

"It identifies cancer at very early stages," she said, noting that it works particularly well for finding lung, breast and colon cancer.

This project was the first time the two siblings worked together, and Lucia said it had advantages and disadvantages over working independently.

"It's fun both ways," she said.

Lucia, who plays the violin, plans to use her share of the scholarship money to help her earn a double major in music and math. Philip plans to study astrophysics.

If they end up being scientists, they will follow in the footsteps of their father, Gabor Mocz, a biochemist and researcher at the University of Hawai'i.

Although their project did not overlap with their father's research, he did provide some guidance.

"He's like our personal hero because he's always there for us," Lucia said.

Reach Treena Shapiro at tshapiro@honoluluadvertiser.com.