ROMA Tool More Sensitive Than RMI to Predict Ovarian Ca

SAN ANTONIO — A novel algorithm has been shown to be more sensitive than a widely used risk of malignancy index for predicting epithelial ovarian cancers in women who present with a pelvic mass or ovarian cyst.

The Risk of Ovarian Malignancy Algorithm (ROMA) stratifies women at high or low risk for epithelial ovarian cancer based on menopausal status and preoperative serum levels of human epididymis protein 4 (HE4) and cancer antigen 125 (CA 125). The algorithm correctly classified 94% of women with epithelial ovarian cancer in a prospective, double-blind, multicenter trial with 457 evaluable patients, researchers said (Gynecol. Oncol. 2009;112:40–6).

A new secondary analysis of trial data comparing patients with benign disease and all stages of epithelial ovarian cancer determined ROMA's sensitivity to be 94.3%, vs. 83.7% for the risk of malignancy index (RMI), when specificity for both was set at 75%. ROMA also was more sensitive than RMI in a comparison of patients with benign disease, tumors with a low potential for malignancy, and epithelial ovarian cancer (89% vs. 80.7%).

“This tool can be used to triage patients to physicians and centers that are experienced in the care and management of patients with ovarian cancer,” Dr. Richard G. Moore said at the annual meeting of the Society of Gynecologic Oncologists. High-risk women should be referred to gynecologic oncologists and centers that have been shown to treat ovarian cancer with better survival outcomes and less morbidity, Dr. Moore said.

The investigators do not see ROMA as replacing the Society of Gynecologic Oncologists/American College of Obstetricians and Gynecologists referral guidelines for pelvic masses, he said, adding that they would like to see ROMA incorporated into the guidelines.

“Clinical findings and impressions are very important, but I think these markers can really help us to triage these patients,” said Dr. Moore of Women and Infants Hospital and Brown University, Providence, R.I. Gynecologists now operate on fewer than half of women with ovarian cancers, he said.

The investigators compared ROMA to RMI because the latter is a validated, well-accepted tool currently in use, he said in an interview. The RMI is based on menopausal status, CA 125 levels, and ultrasound scores of 0–5.

ROMA has one formula for premenopausal and another for postmenopausal women. Both include HE4 and CA 125 levels, but the premenopausal formula weighs HE4 more heavily. “In premenopausal patients, there are many benign diseases that cause elevated CA 125, and there is no cancer,” Dr. Moore said.

The algorithm was based on pooled data from a pilot study at Women and Infants Hospital and a retrospective case-control study at Massachusetts General Hospital, Boston. The prospective trial enrolled 566 women who presented at 12 centers with pelvic masses that were documented on imaging and for which surgery was planned.

Three independent reviewers assigned an RMI score for each evaluable patient based on a preoperative ultrasound, CT, or MRI scan. Patients were included if at least two reviewers agreed on the imaging score; correlation between reviewers was 78.4%. They were blinded to tumor marker values and pathology.

The final population included 212 premenopausal and 245 postmenopausal women. All told, 123 women had epithelial ovarian cancers (80 of which were stage III), 22 had tumors with a low potential for malignancy, and 312 had benign disease.

The investigators did not report sensitivity by histology, but Dr. Moore said it was close to or at 100% in all but mucinous tumors. ROMA was much less sensitive in mucinous tumors, identifying only about half of them, he said.

Although ROMA was significantly more sensitive in other comparisons based on tumor stage, Dr. Moore did not report benign vs. stage I results because only 17 patients were stage I. Even with such small numbers, the comparison trended in favor of ROMA, he said.

The prospective trial was supported by Fujirebio Diagnostics Inc. and grants from the National Cancer Institute. Seven authors, including Dr. Moore, served as consultants to and were on the scientific advisory board for Fujirebio.

The ROMA tool correctly classified 94% of women with epithelial ovarian cancer, said Dr. Richard G. Moore, here with Dr. Geralyn Messerlian. ©David Witbeck

SAN ANTONIO — A novel algorithm has been shown to be more sensitive than a widely used risk of malignancy index for predicting epithelial ovarian cancers in women who present with a pelvic mass or ovarian cyst.

The Risk of Ovarian Malignancy Algorithm (ROMA) stratifies women at high or low risk for epithelial ovarian cancer based on menopausal status and preoperative serum levels of human epididymis protein 4 (HE4) and cancer antigen 125 (CA 125). The algorithm correctly classified 94% of women with epithelial ovarian cancer in a prospective, double-blind, multicenter trial with 457 evaluable patients, researchers said (Gynecol. Oncol. 2009;112:40–6).

A new secondary analysis of trial data comparing patients with benign disease and all stages of epithelial ovarian cancer determined ROMA's sensitivity to be 94.3%, vs. 83.7% for the risk of malignancy index (RMI), when specificity for both was set at 75%. ROMA also was more sensitive than RMI in a comparison of patients with benign disease, tumors with a low potential for malignancy, and epithelial ovarian cancer (89% vs. 80.7%).

“This tool can be used to triage patients to physicians and centers that are experienced in the care and management of patients with ovarian cancer,” Dr. Richard G. Moore said at the annual meeting of the Society of Gynecologic Oncologists. High-risk women should be referred to gynecologic oncologists and centers that have been shown to treat ovarian cancer with better survival outcomes and less morbidity, Dr. Moore said.

The investigators do not see ROMA as replacing the Society of Gynecologic Oncologists/American College of Obstetricians and Gynecologists referral guidelines for pelvic masses, he said, adding that they would like to see ROMA incorporated into the guidelines.

“Clinical findings and impressions are very important, but I think these markers can really help us to triage these patients,” said Dr. Moore of Women and Infants Hospital and Brown University, Providence, R.I. Gynecologists now operate on fewer than half of women with ovarian cancers, he said.

The investigators compared ROMA to RMI because the latter is a validated, well-accepted tool currently in use, he said in an interview. The RMI is based on menopausal status, CA 125 levels, and ultrasound scores of 0–5.

ROMA has one formula for premenopausal and another for postmenopausal women. Both include HE4 and CA 125 levels, but the premenopausal formula weighs HE4 more heavily. “In premenopausal patients, there are many benign diseases that cause elevated CA 125, and there is no cancer,” Dr. Moore said.

The algorithm was based on pooled data from a pilot study at Women and Infants Hospital and a retrospective case-control study at Massachusetts General Hospital, Boston. The prospective trial enrolled 566 women who presented at 12 centers with pelvic masses that were documented on imaging and for which surgery was planned.

Three independent reviewers assigned an RMI score for each evaluable patient based on a preoperative ultrasound, CT, or MRI scan. Patients were included if at least two reviewers agreed on the imaging score; correlation between reviewers was 78.4%. They were blinded to tumor marker values and pathology.

The final population included 212 premenopausal and 245 postmenopausal women. All told, 123 women had epithelial ovarian cancers (80 of which were stage III), 22 had tumors with a low potential for malignancy, and 312 had benign disease.

The investigators did not report sensitivity by histology, but Dr. Moore said it was close to or at 100% in all but mucinous tumors. ROMA was much less sensitive in mucinous tumors, identifying only about half of them, he said.

Although ROMA was significantly more sensitive in other comparisons based on tumor stage, Dr. Moore did not report benign vs. stage I results because only 17 patients were stage I. Even with such small numbers, the comparison trended in favor of ROMA, he said.

The prospective trial was supported by Fujirebio Diagnostics Inc. and grants from the National Cancer Institute. Seven authors, including Dr. Moore, served as consultants to and were on the scientific advisory board for Fujirebio.

The ROMA tool correctly classified 94% of women with epithelial ovarian cancer, said Dr. Richard G. Moore, here with Dr. Geralyn Messerlian. ©David Witbeck

SAN ANTONIO — A novel algorithm has been shown to be more sensitive than a widely used risk of malignancy index for predicting epithelial ovarian cancers in women who present with a pelvic mass or ovarian cyst.

The Risk of Ovarian Malignancy Algorithm (ROMA) stratifies women at high or low risk for epithelial ovarian cancer based on menopausal status and preoperative serum levels of human epididymis protein 4 (HE4) and cancer antigen 125 (CA 125). The algorithm correctly classified 94% of women with epithelial ovarian cancer in a prospective, double-blind, multicenter trial with 457 evaluable patients, researchers said (Gynecol. Oncol. 2009;112:40–6).

A new secondary analysis of trial data comparing patients with benign disease and all stages of epithelial ovarian cancer determined ROMA's sensitivity to be 94.3%, vs. 83.7% for the risk of malignancy index (RMI), when specificity for both was set at 75%. ROMA also was more sensitive than RMI in a comparison of patients with benign disease, tumors with a low potential for malignancy, and epithelial ovarian cancer (89% vs. 80.7%).

“This tool can be used to triage patients to physicians and centers that are experienced in the care and management of patients with ovarian cancer,” Dr. Richard G. Moore said at the annual meeting of the Society of Gynecologic Oncologists. High-risk women should be referred to gynecologic oncologists and centers that have been shown to treat ovarian cancer with better survival outcomes and less morbidity, Dr. Moore said.

The investigators do not see ROMA as replacing the Society of Gynecologic Oncologists/American College of Obstetricians and Gynecologists referral guidelines for pelvic masses, he said, adding that they would like to see ROMA incorporated into the guidelines.

“Clinical findings and impressions are very important, but I think these markers can really help us to triage these patients,” said Dr. Moore of Women and Infants Hospital and Brown University, Providence, R.I. Gynecologists now operate on fewer than half of women with ovarian cancers, he said.

The investigators compared ROMA to RMI because the latter is a validated, well-accepted tool currently in use, he said in an interview. The RMI is based on menopausal status, CA 125 levels, and ultrasound scores of 0–5.

ROMA has one formula for premenopausal and another for postmenopausal women. Both include HE4 and CA 125 levels, but the premenopausal formula weighs HE4 more heavily. “In premenopausal patients, there are many benign diseases that cause elevated CA 125, and there is no cancer,” Dr. Moore said.

The algorithm was based on pooled data from a pilot study at Women and Infants Hospital and a retrospective case-control study at Massachusetts General Hospital, Boston. The prospective trial enrolled 566 women who presented at 12 centers with pelvic masses that were documented on imaging and for which surgery was planned.

Three independent reviewers assigned an RMI score for each evaluable patient based on a preoperative ultrasound, CT, or MRI scan. Patients were included if at least two reviewers agreed on the imaging score; correlation between reviewers was 78.4%. They were blinded to tumor marker values and pathology.

The final population included 212 premenopausal and 245 postmenopausal women. All told, 123 women had epithelial ovarian cancers (80 of which were stage III), 22 had tumors with a low potential for malignancy, and 312 had benign disease.

The investigators did not report sensitivity by histology, but Dr. Moore said it was close to or at 100% in all but mucinous tumors. ROMA was much less sensitive in mucinous tumors, identifying only about half of them, he said.

Although ROMA was significantly more sensitive in other comparisons based on tumor stage, Dr. Moore did not report benign vs. stage I results because only 17 patients were stage I. Even with such small numbers, the comparison trended in favor of ROMA, he said.

The prospective trial was supported by Fujirebio Diagnostics Inc. and grants from the National Cancer Institute. Seven authors, including Dr. Moore, served as consultants to and were on the scientific advisory board for Fujirebio.

The ROMA tool correctly classified 94% of women with epithelial ovarian cancer, said Dr. Richard G. Moore, here with Dr. Geralyn Messerlian. ©David Witbeck