"Long before it's in the papers"
January 27, 2015


Test could provide “family tree” of a patient’s own tumors

April 21, 2014
Courtesy of Massachusetts General Hospital
and World Science staff

A new medical test could pro­vide a “family tree” of can­cer­ous tu­mors in­side a pa­tien­t’s body, show­ing which tu­mors “seed­ed” which oth­ers, ac­cord­ing to a re­port.

That could some­day help with treat­ment plan­ning, the de­ve­lop­ers said. For in­stance, the test could iden­ti­fy the orig­i­nal tu­mor in cases where mul­ti­ple tu­mors ex­ist and where “it is am­big­u­ous which one is re­spon­si­ble for seed­ing” newer ones, said Rakesh K. Jain of Mas­sa­chu­setts Gen­er­al Hos­pi­tal, sen­ior au­thor of the re­port.

The abil­ity of can­cer­ous tu­mors to spread to new sites in the body, called me­tas­ta­sis, is­n’t well un­der­stood, ac­cord­ing to bi­ol­o­gists. It’s not easy to tell wheth­er a par­tic­u­lar tu­mor starts to “seed” oth­ers early or late in its own de­vel­op­ment. Nor is it easily de­ter­mined wheth­er a giv­en tu­mor was “seed­ed” by the orig­i­nal tu­mor where the can­cer be­gan, or by oth­er, lat­er tu­mors.

The re­search­ers said they de­vel­oped a quick, sim­ple test that can re­veal the “fam­i­ly” rela­t­ion­ship­s—which are really ev­o­lu­tion­ary rela­t­ion­ships—among a pa­tien­t’s tu­mors. With such a “family tree,” said study co-au­thor Kamila Naxe­rova, al­so at Mas­sa­chu­setts Gen­er­al, “we could de­ter­mine how dif­fer­ent tu­mors are re­lat­ed to each oth­er and re­con­struct how the can­cer evolved.” 

“Usually that would re­quire ex­ten­sive ge­net­ic anal­y­sis with com­plex se­quenc­ing meth­ods,” but the new meth­od is much more prac­ti­cal, she added. The pa­per is pub­lished this week in the early edi­tion of the re­search jour­nal Pro­ceed­ings of the Na­tio­n­al Aca­de­my of Sci­en­ces. 

Bi­ol­o­gists are only be­gin­ning to in­ves­t­i­gate the ex­tent and im­por­tance of ge­net­ic dif­fer­ences among tu­mor cells, she said. Dif­fer­ences can oc­cur both among dif­fer­ent tu­mors, and with­in a sin­gle tu­mor.

Two dif­fer­ent “mod­els” of me­tas­ta­sis are prev­a­lent, the re­search­ers ex­plained. In one, an orig­i­nal tu­mor “plants” new cells, called metas­tases, else­where in the body late in its own de­vel­op­ment. This sce­nar­i­o pre­dicts only small ge­net­ic dif­fer­ences among dif­fer­ent tu­mors, since they don’t spend a long time de­vel­op­ing apart. In the oth­er mod­el, me­tas­ta­sis oc­curs early in tu­mor de­vel­op­ment, so larg­er ge­net­ic dif­fer­ences be­tween the cells in dif­fer­ent tu­mors are ex­pected. 

Some stud­ies sug­gest the two mod­els apply to dif­fer­ent types of can­cer, but lit­tle hard in­forma­t­ion is avail­a­ble. The new ap­proach fo­cus­es on small ar­eas of the hu­man ge­nome, called “polygua­nine (poly-G) re­peats,” that are par­tic­u­larly sus­cep­ti­ble to muta­t­ion, es­pe­cially dur­ing cell di­vi­sion. These muta­t­ions don’t di­rectly re­late to how tu­mors de­vel­op, but can re­veal the “lin­eages” of dif­fer­ent tu­mor cells in rela­t­ion to each oth­er, ac­cord­ing to the re­search­ers.

They adapted Poly-G re­peat anal­y­sis, in­i­tially de­vel­oped to study rela­t­ion­ships be­tween cells in mice, to study hu­man can­cer. An­a­lyz­ing the poly-G pro­files of pri­ma­ry and met­a­stat­ic co­lon can­cer sam­ples from 22 pa­tients in­di­cat­ed that how the pri­ma­ry and met­a­stat­ic tu­mors re­lat­ed to each oth­er was dif­fer­ent for each pa­tient.

“We found that there are sev­er­al paths” that the rela­t­ion­ships can fol­low in can­cer, said Naxe­rova. The re­search­ers are now using the ap­proach to study “larg­er num­bers of pa­tients,” she added; it’s “fast and in­ex­pen­sive and should be ap­pli­ca­ble to oth­er types of tu­mors than co­lon can­cer.”

* * *

Send us a comment on this story, or send it to a friend

Sign up for

On Home Page         


  • St­ar found to have lit­tle plan­ets over twice as old as our own

  • “Kind­ness curricu­lum” may bo­ost suc­cess in pre­schoolers


  • Smart­er mice with a “hum­anized” gene?

  • Was black­mail essen­tial for marr­iage to evolve?

  • Plu­to has even cold­er “twin” of sim­ilar size, studies find

  • Could simple an­ger have taught people to coop­erate?


  • F­rog said to de­scribe its home through song

  • Even r­ats will lend a help­ing paw: study

  • D­rug may undo aging-assoc­iated brain changes in ani­mals

A new test could provide a “family tree” of cancerous tumors inside a patient’s body, showing which tumors “seeded” which others, according to a report. That could someday help with treatment planning, scientists said. For instance, the test could identify the original tumor in cases where multiple tumors exist and where “it is ambiguous which one is responsible for seeding” newer ones, said Rakesh K. Jain of Massachusetts General Hospital, senior author of the report. The ability of cancerous tumors to spread to new sites in the body, called metastasis, isn’t well understood, according to biologists. It’s not easy to tell whether a particular tumor starts to “seed” others early or late in its own development. Nor is it easily determined whether a given tumor was “seeded” by the original tumor where the cancer began, or by other, later tumors. The researchers said they developed a quick, simple test that can reveal the “family” relationships—which are really evolutionary relationships—among a patient’s tumors. With such a “family tree,” said study co-author Kamila Naxerova, also at Massachusetts General, “we could determine how different tumors are related to each other and reconstruct how the cancer evolved.” “Usually that would require extensive genetic analysis with complex sequencing methods,” but the new method is much more practical, she added. The paper is published this week in the early edition of the research journal PNAS. Biologists are only beginning to investigate the extent and importance of genetic differences among tumor cells, she said. Differences can occur both among different tumors, and within a single tumor. Two different “models” of metastasis are prevalent, the researchers explained. In one, an original tumor “plants” new cells, called metastases, elsewhere in the body late in its own development. This scenario predicts only small genetic differences among different tumors, since they don’t spend a long time developing apart. In the other model, metastasis occurs early in tumor development, so larger genetic differences between the cells in different tumors are expected. Some studies suggest the two models apply to different types of cancer, but little hard information is available. The new approach focuses on small areas of the human genome, called “polyguanine (poly-G) repeats,” that are particularly susceptible to mutation, especially during cell division. These mutations don’t directly relate to how tumors develop, but can reveal the “lineages” of different tumor cells in relation to each other, according to the researchers. They adapted Poly-G repeat analysis, initially developed to study relationships between cells in mice, to study human cancer. Analyzing the poly-G profiles of primary and metastatic colon cancer samples from 22 patients indicated that how the primary and metastatic tumors related to each other was different for each patient. “We found that there are several paths” that the cell’s relationships can follow in cancer, said Naxerova. The researchers are now applying the approach to study cancer in “larger numbers of patients,” she added. “The method is fast and inexpensive and should be applicable to other types of tumors than colon cancer.”