Examination of the breast by mammography provides a sensitive method for early detection of non-palpable malignant tumours. Despite the fact that many breast cancers have been developing for some time before they can be detected by any method, earlier detection can improve the stage at which cancers are diagnosed and can reduce the absolute mortality from breast cancer by 30 to 40 per cent. The mammogram is the most important method of breast examination for detecting occult malignancy in asymptomatic women. Cancers detected by mammography alone are more likely to be at an earlier stage than those detected by physical examination or because of symptoms, with a correspondingly more favourable prognosis.

Mammography is unique in its ability to detect breast lesions that may represent cancers, but it is inaccurate in differentiating benign from malignant disease. Unfortunately, patterns of tumour growth and the breasts' response to them are frequently non-specific. Benign and malignant lesions may have similar morphology: characteristics such as smooth, well defined margins are statistically more common in benign lesions, but some cancers are sharply marginated. Even the classic lucent zone or ‘halo sign’ surrounding circumscribed masses does not guarantee a benign process. At the other end of the spectrum the spiculated architectural distortion of a radial scar, or area of fat necrosis is indistinguishable from the ‘characteristic’ morphological changes produced by breast cancer. Microcalcifications associated with cancer are frequently indistinguishable from those produced by benign processes.

Excision biopsy must therefore be carried out to establish a histological diagnosis on all suspicious non-palpable lesions demonstrated on screening mammography. Since 70 to 80 per cent of these anomalities will be benign upon biopsy examination radiologists must provide safe, accurate guidance for the surgeon to ensure that the lesion in question can be precisely excised with minimum cosmetic defect.

Localization for biopsy examination cannot be successful unless the area in question can be demonstrated in three dimensions. Some lesions will be visible in one projection but not seen in a second. Most lesions are visible in the mediolateral view, but because of their proximity to the chest wall may require special rotated craniocaudal views to triangulate the lesion completely. If these manoeuvres fail, ultrasound examination or computer tomography (CT) can be used.

Since the majority of these lesions are benign, excision of a quadrant as a method of biopsy is contraindicated, not only because it is inaccurate, but because it also unnecessarily removes a large volume of tissue, leading to an unacceptable cosmetic result.

The goal of accurate preoperative localization is the positioning of a needle within 5 mm of the lesion, if not directly through it. Many surgeons prefer that a guide be placed in an anteroposterior direction to facilitate surgery as this orientation parallels their operative approach. However, guides cannot be positioned from the front of the breast and directed back toward the chest wall under direct visualization, as mammograms can only be obtained orthogonal to this direction. There are many who appear to be quite skilled at this technique, although it generally requires repositioning of the guide to achieve 5 mm accuracy in placement. If an anteroposterior needle insertion is used it should be done with caution since needles introduced towards the chest wall can be inserted into the pectoralis muscles, or may pass into the pleural space, lung, or mediastinum. There have been instances of pneumothoraces caused by introducing needles in this direction and wires for localization placed in this fashion that were too short for the depth of the lesion have retracted into the breast when the patient sat up. A few wires have not only retracted completely into the breast, but when not retrieved were massaged by body motion into other anatomical areas.

Such complications are totally avoidable if placement of guidewires is carried out using paths of insertion parallel to the chest wall. The parallel approach is most accurate because the breast can be held firmly in the mammographic compression system during the procedure, with the lesion under direct visualization, allowing the exact relationship of the needle to the lesion to be determined.

Preoperative localization of cancers does not increase the risk of local recurrence when an approach parallel to the chest wall is used.

Although the spot method, or injection of a vital dye in the vicinity of the lesion to stain the tissue has been successfully used to guide surgery, at the Massachusetts General Hospital a needle with a spring hook guidewire introduced parallel to the chest wall is used routinely. This configuration permits the hook to be completely contained in and pass through the needle lumen reforming to hook into the tissue when it is released from the needle tip. The introducing needle can thus be optimally positioned prior to engaging the hook. The guidewire can be afterloaded to anchor in the tissue. The guidewire also incorporates a thick segment just proximal to the hook. This area is positioned by the radiologist in proximity to the lesion so that the surgeon is alerted, during dissection of the wire, to the location of the suspicious area.

Most mammographic units have compression devices with openings that permit the breast to be held within the machine, while affording access to the volume of tissue into which the needle is to be placed. The procedure is begun by evaluating the location of the lesion on the screening mammograms. If the lesion is closest to the medial surface of the breast, initial needle positioning is accomplished by a mediolateral approach, inserting the needle from the medial skin surface. If the lesion is close to the lateral surface of the breast, a lateromedial approach is used. A lesion near the top of the breast is localized from the craniocaudal projection, while one at the bottom of the breast can be localized using a caudocranial projection.

Figure 1(a) 840 demonstrates the localization of a lesion in the medial aspect of the right breast. The breast is compressed within the mammographic device so that a preliminary mammogram projects the lesion in the opening in the compression plate. Its co-ordinates are determined by marks around the plate opening.

A needle is introduced in the direction of the X-ray beam, passing through and beyond the suspicious lesion. Local anaesthesia is rarely required. If initial placement is suboptimal, it is possible to withdraw the needle partially, change the insertion angle, and reinsert without repuncturing the skin. A confirmatory mammogram is obtained through the compression plate. Ideally the hub of the needle should be superimposed on the shaft and on the lesion (Fig. 1(a) inset) 840. The patient is then released from the machine with the needle in place, and the gantry is rotated 90° in preparation for depth determination. The fenestrated compression plate is replaced by a small spot compression device. The breast is recompressed, so that the X-ray beam will be perpendicular to the shaft of the needle, and another confirmatory film is made. If necessary, the needle position is adjusted so that the tip is positioned at the level where the hook will be placed. The hookwire is then inserted and the needle removed after a final radiograph showing the completed placement of the wire. Figure 1(b) 840 shows the hookwire being inserted with the breast in craniocaudal position. The wire is then loosely taped to the skin with sterile tapes without traction. The patient leaves the mammography suite for the operating theatre.

Since only 15 to 29 per cent of lesions removed for biopsy examination on the basis of mammographic suspicion alone contain carcinoma, the goal of the surgeon is to obtain a specimen that will yield an unequivocal diagnosis with minimal risk to the patient in terms of complications and of cosmetic deformity. Accurate localization techniques, as described, permit the surgeon to remove a minimal volume core of tissue surrounding the wire guide that is adequate for diagnosis and does not create a breast deformity. The practice of performing a ‘quadrantectomy’ on all patients to obtain margins negative for cancer if the lesion is malignant is to be condemned since this procedure subjects the 80 per cent of patients with benign processes to the unnecessary removal of a large mass of tissue and to a resultant defect in the contour of the breast.

Figure 1(c) 840 represents the patient upon arrival in the surgical suite. The guidewire protrudes from the skin of the breast and the confirmatory mammogram (inset) shows that the thick segment is centred in the lesion.

Local or general anaesthesia is used for the biopsy, which is performed as an outpatient procedure. After skin preparation and draping, a linear skin incision is made directly along the course of the guide (Fig. 1(d)) 840. Dissection is carried down through the subcutaneous fat to the surface of the breast tissue. Meticulous haemostasis is maintained with the electrocoagulation unit.

Figure 1(e) 840 shows the subcutaneous fat retracted and the thick segment of the wire just appearing at the surface of the breast tissue. The biopsy specimen is then obtained by plunging a No. 10 Bard Parker scalpel directly into the glandular tissue parallel to the guide wire, circumscribing the core of tissue to be removed.

As shown in Fig. 1(f) 840, after the first circular incision is made the core and wire are grasped with Allis' forceps and traction applied to the specimen so that it may be amputated at its base.

The specimen with contained wire (Fig. 1(g)) 840 is placed in a plastic bag on ice (to preserve the level of oestrogen receptors) and is sent to the mammography department for radiological confirmation of the presence of the suspicious area. The specimen should be compressed in the radiographic device so that masses as well as calcifications can be seen. The specimen is then sent to the pathology department.

The diagnosis being confirmed (Fig. 2) 841, haemostasis is completed. The biopsy site is closed with several interrupted absorbable sutures, and the skin is closed with a subcuticular suture of absorbable material.

Basset LW, Gold RH, Cove C. Mammographic spectrum of traumatic fat necrosis: the fallibility of ‘pathognomonic’ signs of carcinoma. Am J Radiol 1978; 130: 119–22.

Brestal JB, Jones PA. Transgression of localizing wire into the pleural cavity prior to mammography. Br J Radiol 1981; 54: 139–40.

Cohen MI, Matthies HJ, Mintyer RA, Keen ME, Murad T. Indurative mastopathies: a cause of false-positive mammograms. Radiology 1985; 155: 69–71.

Collins VP, Loeffler RK, Tivey H. Observations on growth rates of human tumors. Cancer 1956; 9: 988–1000.

Gallagher WJ, Cardenosa G, Rubens JR, McCarthy JA, Kopans DB. Minimal-volume excision of nonpalpable breast lesions. Am J Radiol 1989; 153: 957–61.

Kopans DB, Meyer JE, Lindfors KK, Bucchianeri SS. Breast sonographgy to guide aspiration of cysts and preoperative localization of occult breast lesions. Am J Radiol 1984; 143: 489–92.

Myer E, Kopans DB, Stomper PC, Lindfors KK. Occult breast abnormalities: percutaneous preoperative needle localization. Radiology 1984; 150: 335–7.

Rosenberg AL, Schwartz GF, Feig SA, Patchefsky AS. Clinically occult breast lesions: localization and significance. Radiology 1987; 162: 167–70.

Shapiro S, Venet W, Venet L, Roesser R. Ten to fourteen-year effect of screening on breast cancer mortality. J Natl Cancer Inst 1982; 69: 349–55.

Swann CA, Kopans DB, Koerner FC, McCarthy KA, White G, Hall D. The halo sign and malignant breast lesions. Am J Roentgenol 1987; 149: 1145–7.

Tabar L, Fayerberg CJG, Gad A, et al. Reduction in mortality from breast cancer after mass screening with mammography. Lancet 1985; i: 829–32