It was right out of the ending of the first Indiana Jones movie: valuable artifacts stored away and forgotten for decades in a government warehouse.

The value of these relics is not monetary, but their eventual rediscovery is expected to prove to be invaluable to some families of U.S. service members still missing from the 1950-53 Korean War.

The Joint POW/MIA Accounting Command over the past six years retrieved 14,200 chest radiographs (X-rays) representing about 6,400 GIs missing from Korea, and the data have become key in the identification of Korea War veterans buried as "unknowns" at the National Memorial Cemetery of the Pacific at Punchbowl, officials said.

The accounting command, headquartered at Joint Base Pearl Harbor-Hickam, has started using collarbone and vertebrae X-rays taken when a service member was living, and comparing them with remains previously buried at Punchbowl, to make identifications.

The clavicle and spinal comparisons are among several lab advances made in the past several years that should accelerate the identification of some of the 867 Korean War "unknowns" at Punchbowl, said the command, known as JPAC.

First, JPAC had to get hold of the old X-rays.

John Byrd, director of JPAC’s Central Identification Laboratory, said he began searching for them back in about 2001 or 2002.

"I became aware of them because I read an old paper by an anthropologist who had used those radiographs for a study, and so I thought, ‘Man, those would be a treasure trove for us,’"

JPAC members "kept asking the (National Personnel Records Center) in St. Louis, ‘Where are these things?’ And nobody knew,"

In 2005 the center decided to destroy the X-rays because they were emitting noxious gases, and to recover the silver content.

"Somebody luckily remembered that JPAC had been looking for these things and contacted us and said, ‘You still want these?’"

JPAC, which is tasked with investigating, recovering and identifying missing American war dead, now has chest X-rays for about 75 percent of the Korean War missing, or about 6,400 individuals, officials said.

The X-rays, taken for tuberculosis screening, are particularly useful because DNA to make an identification — a preferred method — can’t be extracted from the Punchbowl unknowns.

According to the Congressional Research Service, the 867 unknowns came from two sources: 416 were turned over by the North Koreans following the armistice, and 451 were from U.S. Graves Registration recovery efforts after the war.

At Kokura, Japan, all the remains were processed by a U.S. Army unit using formaldehyde as a preservative, which damaged the DNA sequence. Since DNA testing did not begin until 40 years later in the mid-1990s, those responsible for preserving the remains were probably unaware of the damage, the report said.

The Congressional Research Service looked at the status of the Punchbowl unknowns at the request of U.S. Sen. Dick Lugar, R-Ind.

Most of the Punchbowl burials took place on Memorial Day 1956. Since then 22 had been disinterred, and of those, one was transferred to the Tomb of the Unknowns at Arlington National Cemetery, 13 had been identified since 1999, and eight were being processed for identification by JPAC, the research service said in a Sept. 27 report.

JPAC provided a more up-to-date snapshot: 11 disinterred this year alone, with six of those identified and three other remains undergoing "final" assurance reviews, the command said.

For 2012, JPAC said it has an even more ambitious goal of 30 Punchbowl unknown disinterments. As many as 400 sets of Punchbowl Korean War unknowns may be able to be tested for identification given current technologies — the most important of which is the chest radiographic comparison, JPAC spokesman Maj. Ramon Osorio said.

The fewer number of disinterments and identifications in the past reflects the "careful approach we have taken over the years," he said.

"Per current policy, we only exhume remains that we believe can be identified in relatively short order," Osorio said.

The new emphasis also comes with increased pressure on JPAC to increase identifications as World War II and Korean War families seeking a return of lost loved ones themselves start to die off.

The clavicle and vertebrae comparisons are now viewed as key potential components of an identification.

Carl Stephan, a JPAC forensic anthropologist, said the command has been working on radiograph comparison for about seven years, but the past three years "is where we’ve really been able to home in on the methods and improve them and test them and make sure they are reliable and then start applying them to the cases that are coming through."

There are "well over dozens" of unique characteristics that can be examined on the relatively short clavicles alone, from their overall shape to ridges and spurs, Stephan said.

With a single clavicle and a clear radiograph, "we can make an identification just off the one bone," said Stephan, who developed JPAC’s standard operating procedures for making the X-ray comparisons.

In a validation test with results published in the spring in the Journal of Forensic Sciences, Stephan said he and another individual were able to correctly pick out chest radiograph matches in a field of 1,000 individuals.

Stephan also has been working to come up with searchable algorithms so a computer can make a match in seconds.

"That’s very pioneering and very difficult to do," said Byrd, the lab director. “He’s been at it a couple of years, and we have partnered with scientists at other federal labs to try to develop these algorithms. I would say we’re probably within about eight months of having a usable program that will search for us.”