G Svane, M.D., Department of Diagnostic Radiology, Karolinska Hospital, Stockholm, Sweden.

Non-palpable breast lesions that are revealed by mammary radiography can be dealt with in various ways (SVANE 1983, SVANE & SILFVERSWÄRD 1983). Needle  biopsy cytology can provide guidance in the choice of treatment, and various methods  have been evolved for such diagnostic examinations of non palpable lesions (MÜHLOW 1974,   BOLMGREN et coll. 1977, NOVAK 1979). The present report deals with non-palpable breast  lesions which were examined with a stereotaxic biopsy instrument (NORDENSTRÖM et coll. 1981)  and were not excised within three months thereafter, but were followed up clinically  and radiographically. The aim of the investigation was to reveal possible 'false' negative  cytologic results.

The needle biopsy was in all cases preceded by routinely  performed mammary radiography, with the compressed breast examined in cranio-caudal,  latero-medial and oblique projections. The aberrant findings were roentgenologically classified as:
(1) Probably benign: a mass with distinct margin with or without  calcifications, calcifications only (e.g. punctate, coarse densities) or an area of in creased   attenuation without radiating structures.
(2) Probably malignant: an irregular mass, an area of increased  attenuation with some radiating structures or a group of irregular microcalcifications.
(3) Malignancy of high probability: an irregular mass with   spicules radiating from the periphery with or without calcifications. All lesions classified as malignancy of high probabilty were immediately operated upon. These cases are reported by SVANE & SILFVERSWÄRD.
Screw needle biopsy technique. The lesions were needled with the  aid of the previously described stereotaxic screw needle biopsy instrument (NORDENSTRÖM  et coIl. 1981). A Rotex Screw Needle Biopsy Instrument ^® was used in all cases  (NORDENSTRÖM 1980). The cellular material was smeared onto glass slides, which as a rule  were air-dried and stained according to the May-Grünwald-Giemsa technique. They were then  cytologically evaluated at the Department of Clinical Cytology, Karolinska Hospital.

The stereotaxically examined material comprised a total of 527  consecutively detected nonpalpable lesions from mammary radiographies of 505 women. The age range of the women was 20 to 89 years, with a mean age of 54.2 and the largest age group   being between 45 and 65 (Table 1). The original radiographic and stereotaxic examinations  were made in the period August 1976 to May 1980.
Of these 527 lesions, 332 were stereotaxically identified and  submitted to screw needle biopsy but were not excised, or were excised more than 3 months after  the biopsy, and 74 could not be identified for biopsy at the stereotaxic examination. The 406  lesions observed on radiography of the breast comprise the material for the present  investigation. At the preceding radiography, 375 of these 406 lesions were roentgenologically regarded as   probably benign and 31 as probably malignant. Twenty-nine of the probably malignant lesions  were reassessed as probably benign at the stereotaxic examination. The remaining 2 lesions  were still regarded as probably malignant but excision was not performed despite recommendation.  The remaining 121 lesions (114 patients) were those already mentioned as being excised  within 3 months of screw needle biopsy (SVANE & SILFVERSWÄRD). The scheme of management of the  patients is surveyed in Fig. 1.
The 317 patients with lesions from which cellular material was  obtained by stereotaxic needle biopsy and the 74 patients, who stereoradiographically proved to   have no lesions were followed up for periods ranging from 2 to 56 months. The length of  follow-up was largely dependent on the clinical course and the degree of malignancy suggested at the  original mammary radiography, and also on observations at subsequent  radiographies. The purpose of the clinical and radiographic surveillance was to reveal possible "false"  negative cytologic results.

The 332 lesions which were identified in the breasts at  stereotaxic screw needle biopsy are subgrouped in Fig. 2. Of the 9 patients not being followed up  (a), 5 died (of unrelated diseases) during the observation period, 3 left the hospital's catchment   area and one failed to attend for re-examination. Of the 31 lesions in subgroup (b), 18 proved to  be cysts, from which fluid was aspirated and which thereafter disappeared or diminished  appreciably in firmness and size. Five lesions consisted of normal lymph glands and 8 were cytologically  considered to be fibroadenomas. The size of these 31 lesions, the largest  radiographic diameter being measured, was 5 mm or less in four cases, while ten were 6-10 mm, eight  11-15 mm, six 16-20 mm and three larger than 20 mm.
Nine lesions (subgroup c) were excised more than 3 months after   screw needle biopsy. The reasons for excision were enlargement at subsequent mammary  radiographies (4 lesions) or that excision, for other reasons, was considered to be more  appropriate than continued clinical follow-up. The cytologic and histopathologic findings in these 9  lesions are presented in Table 2, which also demonstrates the maximum roentgenographic diameter  of the malignant lesions.
The radiographic report from 2 of the histopathologically  malignant lesions stated probable malignancy, while 2 were regarded as probably benign. One of the  first two lesions was excised after slightly more than 3 months, without further radiography,   and the other after 7 months, because of enlargement at a radiographic follow-up. One of the  two probably benign lesions was excised after 15 months, also because of enlargement at a  follow-up radiography. The remaining histopathologically malignant lesion was smaller on  films taken 6 months after the stereotaxic biopsy. After one and a half year, however,  radiography showed within the same quadrant of the breast a lesion of malignancy of high probability  and there was palpable thickening of tissue in this area. Conventional aspiration biopsy  now yielded malignant cells. It was not possible to decide if the malignant tumor had arisen  after the stereotaxic biopsy, or if the same lesion had been needled on both occasions. The  stereotaxic investigation could thus have given a "false" negative cytologic diagnosis. All 5 of the   histopathologically benign tumors (Table 2) had been classified at mammary radiography as  probably benign and the stereotaxic screw needle biopsy had yielded benign cellular  material.
At subsequent radiographies 120 lesions (subgroup d, Fig. 2) had  diminished in size (56) or disappeared (64). The radiographic and cytologic diagnoses of  these 120 lesions are shown in Table 3 together with the follow-up period and the diameter of  the lesions. The 2 lesions which originally were considered to demonstrate atypical epithelium  could no longer be found at subsequent radiographies.
In 146 lesions there were no changes in size or appearance during  the follow-up period (subgroup e, Fig. 2). Data concerning these lesions are presented   in Table 4. All 3 lesions with an atypical cytologic appearance were observed for 25 months  after the initial mammary radiography.
Three patients failed to co-operate in the biopsy procedure,  which therefore was incorrectly performed (subgroup f, Fig. 2).
The initial biopsy yielded insufficient material for diagnosis of  10 of the lesions (subgroup g, Fig. 2). Subsequent biopsy in 3 of them demonstrated benign   epithelium. All of these patients were radiographically followed up until the observed area was  considered not to differ from normal breast tissue or to contain a benign lesion. The  observation periods were as follows: 12 months 2 lesions, 13-18 months 2 lesions, 19-24 months 3 lesions,  25-36 months 2 lesions, and more than 3 years in the remaining case. The radiographic assessment of all 10 lesions was probably benign. Three of them were less than 5 mm at their  largest diameter, while four were 6-10 mm, two 11-15 mm and one 16-20 mm.
In screw needle biopsy of 4 well defined lesions the needle  displaced the target tissue and the yield was non-representative (subgroup h, Fig. 2). All 4 biopsies  were done at the beginning of the examination period and the lesions presented firm resistance  to the needle. The largest roentgenographic diameters were 3, 8, 15 and 20 mm, respectively.  They were radiographically followed up for 7, 9, 21 and 25 months without discernible  changes in the lesions. With increasing experience of the biopsy technique, needling of such   lesions ceased to be a problem. It is possible to advance the screw needle carefully into the  lesion while constantly rotating the needle in both directions. In this way the target does not  deviate, even if it offers firm resistance. The 74 lesions in which stereotaxic examination could  not confirm the radiographic observations are presented in Fig. 3. In addition to the 2  lesions that proved to be abnormalities in the skin, there were 2 that could not be reached by the  stereotaxic biopsy procedure. One of these latter cases consisted of calcifications located   peripherally in the breast, and the other one was a possible recurrence in a small residue of breast tissue  after radical mastectomy. Of the other 70 lesions not confirmed at stereotaxic examination, 68  were radiographically followed up and these examinations revealed a lesion in only 7 cases. In 4 of  them the lesions were too diffuse to enable a representative puncture to be chosen. The  other 3 were calcifications, in one case intravascular. The calcified lesions were not excised, but   were observed for 24 to 27 months without any evidence of change in appearance.