The cost-effectiveness of fine-needle aspiration cytology and 14-gauge core needle biopsy compared with open surgical biopsy in the diagnosis of breast carcinoma
Judicious utilization of fine-needle aspiration cytology (FNAC) and 14-gauge core needle biopsy (CB) theoretically should result in greater accuracy in breast carcinoma diagnosis and fewer unnecessary open surgical biopsies (OSBs), thus lowering health care costs.
In 1995 in Rochester, New York, the ratio of open surgical breast biopsies per each verified breast carcinoma (OSB/Ca) in a freestanding breast clinic (EWBC) was compared with the OSB/Ca ratio of all physicians in the remainder of the city. The EWBC differs from all other diagnostic facilities in Rochester in that it routinely performs FNAC and CB.
The EWBC recommended 462 OSBs resulting in 310 verified carcinomas, for a OSB/Ca ratio of 1.5. The physicians in the remainder of the city recommended 2036 OSBs resulting in 513 verified carcinomas, for a OSB/Ca ratio of 4.0. If the EWBC OSB/Ca ratio had been identical to the remainder of the city, the number of extra OSBs recommended by the clinic would have been 778, resulting in an additional cost of $1,712,082. When the added cost of the 2594 FNACs ($256,285) and 403 CBs ($252,278) performed by the clinic was subtracted from the $1,712,082, the freestanding breast clinic cost savings was $1,203,519. The lymph node metastasis rate of 19% for the breast carcinomas diagnosed in clinic patients was identical to that of the women with breast carcinoma in the remainder of the city.
Utilization of FNAC and CB allows radiologists to lower their OSB/Ca ratio without sacrificing early detection. In this study, these less expensive procedures result in lowered medical costs for the health care system. Cancer 1998;82:1867-73. © 1998 American Cancer Society.
The leading cause of death due to cancer in women age < 65 years is breast carcinoma. In the U. S. in 1995, > 180,000 cases of breast carcinoma were diagnosed, but it is estimated that it was necessary to perform 5 open surgical biopsies (OSBs) for every case of breast carcinoma that was diagnosed.
The Elizabeth Wende Breast Clinic (EWBC), founded in 1976 in Rochester, New York, is a freestanding clinic utilizing imaging and interventional diagnostic procedures dedicated solely to the diagnosis of breast carcinoma. The only physicians at the EWBC are diagnostic radiologists. Approximately 75% of the patients live in the Rochester area and 25% of the patients are referred to the clinic from western New York regions outside the Rochester area. The clinic, as well as the majority of other diagnostic radiology facilities in the Rochester area, performs screening and diagnostic mammograms and utilizes ultrasonography on patients with radiographic abnormalities and clinical problems. However, the EWBC is unlike all other Rochester diagnostic radiology facilities in that its diagnostic armamentarium also includes routine use of fine-needle aspiration cytology (FNAC), as well as 14-gauge core needle biopsy (CB) performed under stereotactic, ultrasonographic, and clinical guidance.1-3 The physicians at the EWBC have developed a systematic method to evaluate these patients using flow charts refined over the past 22 years.4 If, after evaluation, a physician at the EWBC believes that the patient most likely has breast carcinoma, the patient's lesion then falls above the physician's OSB threshold and therefore warrants surgical removal. All lesions below the OSB threshold are considered to be most likely benign by the physician. Each physician has her own OSB threshold for each lesion, depending on its clinical, mammographic, and ultrasonographic characteristics.
Several years after acquiring two prone-table stereotaxic core needle biopsy units (Mammotest/Mammovision system; Fischer Imaging, Denver, CO, and Stereoguide; LoRad, Danbury, CT), the EWBC recognized that although it was recommending fewer OSBs, the FNAC/CB combination was detecting earlier tumors in patients with subsuspicious lesions. Therefore the clinic reviewed its 1995 data, then focused on the subset of patients from the city of Rochester (75% of all patients) and compared its OSB/ breast carcinoma (Ca) ratio with breast biopsy data from the remainder of the city of Rochester to determine whether the clinic's use of FNAC and CB was cost-effective.
MATERIALS AND METHODS
The physicians at the EWBC perform FNAC5 and CB for three reasons: 1) to verify a suspected carcinoma so the surgeon has the option to proceed directly to lymph node dissection and either mastectomy or lumpectomy, thus potentially reducing the number of open surgical procedures from two to one; 2) as a corroborative cross-check, once it is decided that the mammographic, ultrasonographic, and clinical findings are within normal limits or represent a benign condition such as a fibroadenoma, and 3) to prevent an unnecessary biopsy on a benign lesion, such as a lipoma that is palpable as a discrete mass.
When it is decided that a patient most likely has breast carcinoma, FNAC is performed if the lesion is palpable. Otherwise, CB is performed. CB also is performed if FNAC does not verify a definite carcinoma. FNAC is utilized first, whenever possible, because it is less invasive and a less expensive way to diagnose carcinoma than CB.
If a physician at the EWBC believes that a clinical or radiographic aberration is benign, FNAC is performed as a cross-check on those lesions that clinically are palpable or ultrasonographically visible. CB is obtained when the cytology laboratory interprets the FNAC as "atypical cells, recommend biopsy," or when FNAC cannot be performed because the abnormality is nonpalpable and nonvisible ultrasonographically. Because of the greater accuracy of diagnosis,6-8 the radiologists at the clinic also proceed directly to CB if a lesion is extremely close to, but nevertheless below, their threshold for recommending an OSB.
The EWBC acquired a list of the pathology reports on all women who underwent OSBs of the breast in 1995 from each of the eight hospitals in Rochester. To protect each patient's privacy, the only identifying data were the surgical accession numbers. From this list, the EWBC tabulated the number of diagnosed breast carcinomas and benign conditions. Biopsies on breast skin lesions, acute abscesses, and plastic surgery procedures were not included. Because the EWBC routinely obtains a pathology report on every one of its patients who is biopsied, the clinic was able to subtract its patients with histopathologically diagnosed breast carcinoma and benign conditions from each hospital's list. The remainder of the OSBs therefore were performed on Rochester patients for whom OSB was not recommended by the EWBC clinic. In this manner the EWBC OSB/Ca ratio could be compared with the OSB/Ca ratio of those women in Rochester whose biopsies were not recommended by the EWBC.
The average reimbursement for each procedure performed at the EWBC was determined by dividing the number of each procedure performed into the sum total reimbursement received for that procedure from all insurance carriers and patients. The reimbursement amounts for cytologic and histopathologic interpretations were obtained from Strong Memorial Hospital, where the specimens were submitted.
The cost of each OSB was obtained by contacting a representative surgeon from each of the hospitals in Rochester, adding total reimbursement from insurance carriers and patients, and dividing this by the total number of OSBs performed. We then calculated the average reimbursement amounts for all of the surgeons. In the same way, the average reimbursement for the operating room and anesthesiologist were obtained.
A short term follow-up ultrasound or unilateral mammogram was performed in < 1 year on some of the patients who underwent FNAC, CB, or OSB. These costs also were included.
The EWBC did not include mammography and ultrasonography at the time of the patients' initial visits to the EWBC in its cost-effectiveness analysis because we were unable to obtain this information from the other physicians in Rochester for comparison, but the EWBC costs most likely were canceled out by even higher costs incurred by patients evaluated elsewhere in the community. For example, all the patients who the EWBC saw for second opinions had undergone mammograms and the majority had undergone sonograms. It was apparent that these patients had been evaluated in the same manner the EWBC radiologists evaluate their patients, with the exception that the outside facilities had not performed FNAC or CB. In fact, there most likely was an increased average cost of evaluating patients elsewhere in Rochester due to surgical consultation expenses, because 258 of the 1527 women (17%) who were referred to the EWBC for second opinions had, in addition to being evaluated at a radiologic facility, already been referred to and had been seen by a surgeon for a screening-detected aberration or clinical problem.
Total EWBC Data
In 1995, the EWBC saw three categories of patients and performed further diagnostic and/or interventional procedures when indicated on: 1) 31,694 women who underwent screening mammograms, 2) 12,423 women with clinical problems such as palpable masses or nipple discharge, and 3) 1527 women referred to the clinic for a second opinion after evaluation at another diagnostic radiologic facility. Of these women, 20% had clinical problems. The remainder had aberrations on their screening mammograms.
All patients evaluated at the EWBC clinic in 1995 were followed from 1 to 2 years until the end of 1996, to determine the number of 1-year-interval missed breast carcinomas. A carcinoma emerged within 1 year of a negative evaluation in 40 women, accounting for 9% of all the carcinomas diagnosed on EWBC patients. Lymph node metastases were present in 20% of these patients. During 1995, a false-negative FNAC was obtained from 50 of 320 patients with breast carcinoma (15.6%). In all but three of these patients, the radiologist was sufficiently concerned regarding the radiographic and/or clinical appearance to proceed to CB or to recommend OSB despite the negative FNAC results. In each of the remaining 3 patients, the false-negative FNAC resulted in a delayed diagnosis of 3 to 4 months. CBs were performed on 28 of the patients whose FNACs were false-negative; these all were interpreted correctly as suspicious of carcinoma or frankly malignant. Of the total of 155 patients with breast carcinoma on whom CB was performed, there were 11 false-negative interpretations (7%). The radiologist was sufficiently concerned regarding the clinical and/or radiographic appearance to recommend OSB despite the negative CB interpretation in all but three of these patients. The diagnosis of breast carcinoma on these three patients was delayed by 4 months, 8 months, and 1 year, respectively. The sensitivity of FNAC was 87% and its specificity was 98%. The sensitivity of CB was 93% and its specificity was 95%.
Of the 385 CBs performed on women without breast carcinoma, there were 15 false-positive interpretations (3.9%). OSBs performed on these women revealed ductal hyperplasia in all cases except for one woman with apocrine metaplasia and one woman with a radial scar. Of the 3155 FNACs performed on women without breast carcinoma, there were 65 false-positive interpretations (2%). Thirty-three fibroadenomas were diagnosed after performing 15 OSBs and 18 CBs. The remaining 32 lesions, diagnosed after 11 CBs and 21 OSBs, primarily were comprised of ductal hyperplasia without atypia (24), apocrine metaplasia (3), papilloma (2), gynecomastia (2), and a radial scar (1). None of the false-positive CBs was performed on women with false-positive FNACs. All the women who had false-positive CBs or FNACs were free of abnormalities 2-3 years later.
Of all the EWBC patients with breast carcinoma, there were four patients for whom the histopathologic interpretations of their OSBs were benign (1% false-negative). In one of these patients, the OSB did not yield the suspicious lesion. Repeat OSB resulted in successful removal of the carcinoma. Repeat OSBs were not necessary for the other three patients, because the histopathologic diagnoses were changed to carcinoma on review and correlation with the CB results. We were unable to obtain the false-negative rate incurred by all other physicians in the city of Rochester.
During 1995, FNAC performed as a cross-check on "probably benign" lesions detected 11 breast carcinomas. CB performed on "probably benign" lesions detected 38 breast carcinomas. Of these 49 FNAC- and CB-detected lesions, only 3 (6%) had metastatic lymph node involvement.
When the EWBC unequivocally established the diagnosis of breast carcinoma with FNAC or CB, the surgeons performed only one open surgical procedure (treatment) on 73% of the patients (Table 1). However, if the EWBC did not perform FNAC or CB and therefore did not establish the diagnosis of breast carcinoma, the surgeons found it necessary to perform > 1 open surgical procedure on 71% of the patients (diagnosis then treatment).
Table 1. FNAC/CB Results of the 421 Breast Carcinomas Identified at The Elizabeth Wende Breast Clinic in 1995
|If no FNAC/CB performed||35||10||29%||25||71%|
|If definite breast carcinoma diagnosis obtained with either FNAC and/or CB||268||195||73%||73||27%|
The differences between screening patients and patients with clinical problems are shown in Table 2. There were disproportionately more FNACs performed on patients with clinical problems because the majority of these patients had palpable areas believed to be normal variations. There were disproportionately more CBs performed on patients with screening abnormalities because many of these patients had calcifications or small masses that were neither clinically palpable nor ultrasonographically visible. The OSB/Ca ratio of screening patients was 1.3 compared with a 1.7 ratio of patients with clinical problems. The larger percentage of benign OSBs in women with clinical problems was attributed mostly to OSBs on 59 patients with ductogram abnormalities (mostly papillomas), and 44 patients with fibroadenomas (false-positive FNACs, growth spurts, and patients who preferred removal).
Table 2. The Elizabeth Wende Breast Clinic 1995 Open Surgical Biopsy/Breast Carcinoma Ratios (Screening vs. Clinical Problems)
|No. of patients||31,694||13,950|
|No. of FNAC||636||2,823|
|No. of CB||426||111|
|No. of OSB||313||323|
|No. of diagnosed breast carcinomas||231||190|
|% diagnosed breast carcinomas||0.73%||1.3%|
|No. 1-year-interval missed breast carcinomas||28||11|
|% 1-year-interval missed breast carcinomas||11.1%||5.6%|
The EWBC 1-year-interval missed breast carcinoma rate for screening patients (Table 2) was 11%, which compares favorably with the 11-25% rates described in the literature.9-11 The clinic was unable to compare its 6% 1-year-interval missed breast carcinoma rate in symptomatic patients because we were unable to find any information in the medical literature regarding the missed rate of breast carcinoma in symptomatic patients by radiologists utilizing FNAC and CB.
The advantage of FNAC and CB utilization was reflected in the changed recommendations among many of the 1527 women referred to the EWBC for a second opinion in 1995. Approximately 80% of these women had mammographically detected lesions. The remaining 20% were patients with clinical problems, usually palpable abnormalities. Of the 1107 patients whose mammographic and/or clinical findings were considered benign by outside facilities, 27 carcinomas were diagnosed when the EWBC recommended 46 OSBs after performing 38 FNACs and 27 CBs. The FNACs yielded 17 carcinomas and the CBs yielded 10 carcinomas in these 27 patients. Because of the suspicion of carcinoma OSB was recommended by outside facilities in 420 of the patients referred for a second opinion. The EWBC originally canceled 263 of these OSBs, but performed 62 FNACs and 59 CBs to be certain its opinion was correct. The FNACs yielded 3 carcinomas and the CBs yielded 12 carcinomas. The revised total of canceled OSBs was 248. After performing FNAC and/or CB, the EWBC recommended OSBs on the remaining 172 women, verifying the diagnosis of 120 carcinomas.
The costs of the procedures are summarized in Table 3. In 1995 the EWBC recommended 462 lumpectomies in women from Rochester who were seen at its clinic; of these patients, 310 breast carcinomas were diagnosed for an OSB/Ca ratio of 1.5. The physicians in the remainder of the city recommended 2036 OSBs; of these patients, 513 breast carcinomas were diagnosed for an OSB/Ca ratio of 4.0. If the EWBC OSB/Ca ratio had been 4.0, the clinic would have recommended 778 additional OSBs for a total cost of $1,712,082.
Table 3. Cost Figures for FNAC, CB, and OSB
|$42.00 AvR for procedure x 2594 EWBC patients||$108.948|
|$43.00 AvR for cytopathologic interpretation x 2594 EWBC patients||$111,542|
|Total cost of 2594 EWBC FNAC||$220,490|
|$95.00 AvR for <1 year follow-up ultrasound x 1582 EWBC patients (61% of 2594)||$150,290|
|Total cost of EWBC FNAC plus follow-up||$370,780|
|Offset by $85.00 AvR for procedure x 1347 patients done elsewhere in Rochester||-$114,495|
|Total cost of FNAC incurred above and beyond the rest of the city||$256,285|
|$437.00 AvR for procedure x 403 EWBC patients||$176,111|
|$134.00 AvR for histopathologic interpretation x 403 EWBC patients||$54,002|
|Total cost of 403 CB||$230,113|
|$55.00 AvR for <1 year follow-up unilateral mammogram x 403 EWBC patients||$22,165|
|Total cost of CB plus follow-up||$252,278|
|Total cost FNAC + CB||$508,563|
|$441.00 AvR for surgeon x 778 patients||$343,098|
|$878.00 AvR for operating room x 778 patients||$683,084|
|$667.00 AvR for anesthesiologist for 1 hour x 778 patients||$518,926|
|$134.00 AvR for histopathologic interpretation x 778 patients||$104,252|
|Total cost 778 OSBs||$1,649,360|
|$55.00 AvR for <1 year follow-up unilateral mammogram x 54 patients (7% of 778 patients)||$2,970|
|$194.00 AvR for preoperative wire loc. x 308 (39.6% of 778 patients)||$59,752|
|Total cost of OSB plus follow-up||$1,712,082|
|Total cost 778 OSBs||$1,712,082|
|Minus (-) total cost CB, extra FNAC, and short term follow-up||-$508,563|
|Total savings for 1995||$1,203,519|
During 1995 in Rochester, the EWBC performed 2594 FNACs, for a total cost of $370,780. Elsewhere in Rochester, surgeons and a few radiologists performed 1347 FNACs at a total cost of $114,495; when that amount was subtracted from the EWBC's FNAC costs, the remaining amount ($256,285) represented the EWBC's extra FNAC costs above and beyond that of the remainder of the city. The clinic performed 403 CBs (361 stereotactic biopsies, 30 ultrasonographically guided biopsies, and 12 clinically guided biopsies), for a total cost of $252,278. No CBs were performed elsewhere in Rochester in 1995. When the EWBC's extra costs of the FNACs and CBs were subtracted from the cost of the 778 fewer OSBs recommended by the clinic, the savings during 1995 amounted to $1,203,519.
There most likely were additional savings, but the EWBC was unable to obtain the data to calculate the exact amount. Some of these savings were due to fewer open surgical procedures in the patients for whom the EWBC established the diagnosis of breast carcinoma with FNAC and CB (Table 1). There were other savings because, in contrast to elsewhere in the Rochester community, only a few of the patients whom the EWBC determined to have benign conditions obtained surgical consultations.
The percentage of lymph node metastases in the EWBC patients with diagnosed breast carcinoma was compared with the rate of lymph node metastases in the women with diagnosed breast carcinoma in the remainder of the city of Rochester to determine whether the fewer OSBs recommended by the EWBC resulted in the diagnosis of carcinoma at a later stage. The lymph node metastasis rate of the carcinomas diagnosed by the EWBC was 19%, identical to that of the remainder of the city.
In the years prior to the EWBC's routine utilization of FNAC, many of the clinic's patients with benign clinical problems such as lipomas, fibroadenomas, and hormonally affected glandular tissue underwent OSBs. The clinic's policy of performing a cross-check FNAC on patients with benign conditions now accounts for many of the canceled biopsies on these women. In fact, many of the FNACs performed by surgeons elsewhere in Rochester in 1995 most likely also were performed as a cross-check. If a patient's clinical condition was considered to be benign, the EWBC radiologists performed CB when the FNAC interpretation was "atypical cells, recommend biopsy," or when a palpable abnormality was just below the radiologist's clinical threshold for recommending OSB. However, in 1995, surgeons elsewhere in Rochester performed OSB instead of CB in such situations.
In the U. S., the cost of all additional diagnostic procedures performed to evaluate aberrations observed on screening mammograms (downstream cost) is particularly vexing.12-19 The greatest contribution of CB and FNAC at the EWBC has been to reduce the number of unnecessary OSBs that previously were necessary to obtain a histopathologic diagnosis in women who possibly could have breast carcinoma, but most likely do not.
Ideally, all radiologists would be equally skilled in interpreting a mammogram and everyone would agree on the criteria for biopsying a lesion or following it at short intervals. But radiologists often disagree in their interpretations and recommendations. Howard et al.20 compared ten radiologists who reviewed identical sets of high quality mammograms. These radiologists disagreed in their interpretations of 33% of the films. Even when they agreed, they differed radically in their recommendations for the next steps in managing the associated radiographic problems of 25% of the patients. Some recommended that the patient return at the age-appropriate time for a routine mammogram, some advised monitoring the lesion at short or longer intervals, and others counseled the patient to undergo biopsy.
The literature reports that the OSB/Ca ratio of radiologists can vary from 1.5 to 10.21-33 At one end of the spectrum21 are those radiologists who recommend OSB to as many as ten women for each carcinoma confirmed. These radiologists could perform CB on those patients with the most benign-appearing lesions whom they ordinarily would send for OSB, thus lowering their OSB/Ca ratio.
At the opposite end of the spectrum33 are radiologists who recommend an OSB to only three women for every two carcinomas that are verified pathologically. This group of radiologists could perform FNAC and/or CB on lesions that they ordinarily would not send for OSB, but would monitor at short intervals. In other words, they would perform FNAC and CB on those lesions that are below their threshold for recommending OSB. Therefore, FNAC and CB can detect more early breast carcinomas in women whose mammograms are assessed by the radiologist with a low OSB/Ca ratio.
It might be argued that the expertise in interpreting mammograms at the EWBC accounted for the clinic's fewer number of OSBs recommended for each carcinoma diagnosed. However, it became clear on reviewing the mammograms submitted for second opinion that the interpretations of radiologists in the Rochester community are excellent. The physicians at the clinic believe that if they had not utilized FNAC and/or CB to evaluate these indeterminate lesions, they also would have recommended the same number of OSBs on these patients. Therefore, the EWBC physicians do not believe that their expertise in interpreting mammograms contributed significantly, if at all, to the lower OSB/Ca ratio.
The equal incidence of lymph node metastases in the EWBC patients with breast carcinoma compared with those in the remainder of Rochester strongly suggests that the lower EWBC OSB/Ca ratio does not result in delayed diagnoses.
The utilization of FNAC and CB is a cost-effective way to spare women from unnecessary OSB without compromising breast carcinoma detection. Application of these methods resulted in a cost savings of > $1,000,000 in 1 year.