- Email: [email protected]
The Use of Carbon Suspension as an Adjunct to Wire Localisation of Impalpable Breast Lesions
Clinical Radiology (2002) 57: 937±944 doi:10.1053/crad.2002.0994, available online at http://www.idealibrary.com on
The Use of Carbon Suspension as an Adjunct to Wire Localisation of Impalpable Breast Lesions H . A . M O S S , S . J . B A R T E R , M . NAYA G A M , D. L AW R EN C E , M . P I T TA M *Department of Radiology, Addenbrooke's Hospital, Hills Road, Cambridge, {Departments of Radiology, {Pathology and }Surgery, Beds and Herts Breast Screening Service, Luton and Dunstable Hospital NHS Trust, Lewsey Road, Luton Bedfordshire LU4 0DZ, UK Received: 12 October 2001 Revised: 27 March 2002
Accepted: 5 April 2002
AIM: To determine the accuracy and therapeutic success of localisation of impalpable breast lesions by hookwire with additional lesion marking with carbon suspension to mark screen detected abnormalities requiring surgical excision. MATERIALS AND METHODS: Retrospective review of all breast localisation procedures performed in our unit on women with a screen detected abnormality requiring excision over a 7 year period. RESULTS: One hundred and thirty eight women underwent breast localisation procedures. All of the mammographic abnormalities were excised at the initial surgical procedure. The benign to malignant ratio was 1 : 2. Pre-operative cytology was used to guide the extent of surgical excision, with clear margins in 70 of the 92 patients (75%) with malignancy. Twenty patients had further surgery: mastectomy in 7 and further local excision in 14. The localisation procedure was a therapeutic success in the local excision of malignancy in 73 of the 92 patients (79%) with malignancy. CONCLUSION: This method of localisation biopsy is an accurate technique for surgical excision of mammographically detected impalpable abnormalities. The surgeon is able to choose the site of surgical incision to give the best cosmetic result, the lesion is easier to identify at operation and the con®dence that the abnormality has been excised is improved. Moss H. A. et al. (2002) Clinical Radiology 57, 937±944. # 2002 The Royal College of Radiologists Published by Elsevier Science Ltd. All rights reserved. Key words: carbon, breast biopsy, breast surgery, breast neoplasms, localisation.
INTRODUCTION
The National Health Service Breast Screening Programme has resulted in the detection of numerous impalpable lesions in asymptomatic women. After assessment clinically, with imaging and ®ne needle aspiration cytology (FNAC) or core biopsy, some of these lesions require surgical excision biopsy. Accurate preoperative localisation is essential to enable surgical removal of the area of concern whilst also achieving an optimal cosmetic result and minimum patient morbidity. Guidelines from the British Association of Surgical Oncologists (BASO) advise a maximum biopsy weight of 20 g for lesions not proven to be malignant preoperatively [1]. A variety of localisation techniques have been described using mammographic stereotactic [2±20], ultrasound [7,17,19,21±24] and magnetic resonance imaging guidance [25,26]. Methods of marking the abnormality to guide the surgeon to the lesion include insertion of a hookwire Address for correspondence and guarantor of study: Dr SJ Barter Beds and Herts Breast Screening Service, Luton and Dunstable Hospital NHS Trust, Lewsey Road, Luton, Bedfordshire, LU4 ODZ, UK. Fax: 01234-792106; E-mail: [email protected] 0009-9260/02/$35
[2±20,22], injection of a dye such a methylene blue [5,7,17,27] or carbon [7,8,17,18,24,28,29] or marking the skin [17,21,22]. The most frequently used technique is hookwire insertion into the lesion under imaging control, using stereotactic mammographic guidance [2±20]. The patient is then taken to the operating theatre where the surgeon commonly makes an incision at the site of hookwire insertion and follows the wire down to the hook, which is excised along with the lesion. Problems encountered with this technique include signi®cant rates of failure to excise the lesion in up to 17.9% of patients. Studies have reported success rates of localisation ranging from 82.1% to 93.7% [2,3,5,15,27]. In addition the surgeon must make the incision close to the point of insertion of the wire which may not give the best cosmetic result, and which may result in a long track to the abnormality. Also, specimen radiographs may be dicult to interpret with small and subtle lesions of the kind that are detected by breast screening. In 1993, Svane described the use of carbon suspension for the localisation of impalpable mammographic abnormalities [8]. A 4% aqueous suspension of carbon was injected into the lesion under stereotactic guidance. As the needle was
# 2002 The Royal College of Radiologists Published by Elsevier Science Ltd. All rights reserved.
938
CLINICAL RADIOLOGY
withdrawn, carbon was continuously injected to produce a tract of carbon from the lesion to the skin, where the carbon produced a small mark or tattoo. The surgeon was then able to follow the carbon track to the lesion and excise it. Carbon was found to be ideal for marking since it did not diuse into surrounding tissue and its inert properties meant that the localisation procedure did not have to be carried out immediately prior to surgery. In our unit, we use a combination of a rigid hookwire (Hawkins Type I) and carbon suspension to localise impalpable lesions. We describe this technique, its advantages and the diagnostic and therapeutic results of breast localisation using this method. MATERIALS AND METHOD
We retrospectively reviewed all breast localisation procedures performed on women in the screening population who attended our Breast Screening Service for localisation of screen detected impalpable lesions between October 1991 and December 1998. The clinical and imaging ®ndings, pre-operative cytology, and method of localisation was collected from the Breast Screening database. Histology of the excised specimen, its size, weight and excision margins were obtained from the pathology reports. Any further surgical procedures such as re-excision or mastectomy were noted. Statistical analysis was performed using the MannWhitney two sample tests. A combination of carbon injection and the hookwire technique was used to localise the mammographic abnormality, either using stereotactic control (Siemens Mammomat) or ultrasound guidance (Siemens Sonoline). The skin was anaesthetised with 1% lignocaine and a 21 gauge, 80 mm needle (Guilford needle, Mediplus Ltd) was inserted into the lesion and its position checked with swing views. 0.2±0.4 mls of sterile carbon suspension 40 mg/mls (Apoteket Laboratories) was then injected into the lesion. The needle was then withdrawn and a Hawkin Type I hookwire inserted into the lesion without moving the patient. The position of the hookwire was checked by two mammographic views in orthogonal planes and a diagram made of the needle position in the patient's notes. In 6 patients with segmental microcalci®cation several injections of carbon were made to de®ne the posterior and lateral margins of the mammographic abnormality. If the mammographic abnormality was super®cial, no hookwire was inserted but a carbon track was made from the lesion to the skin surface by constantly injecting a small amount of carbon suspension as the needle was withdrawn, leaving a tattoo on the skin. Surgical excision was performed under general anaesthetic on the same day. As the Hawkins wire is rigid the surgeon is able to palpate its tip and choose the incision site for best approach to the lesion and cosmesis. Using gentle traction on the hookwire, the surgeon dissects down to the tip to locate the carbon. The tissue containing the carbon is then excised and a specimen radiograph taken using mild compression. The result of the specimen radiograph is communicated to the surgeon by the radiologist who performed the localisation procedure to discuss if excision is
adequate. The histology results are reviewed at a weekly multidisciplinary meeting and further management discussed. (Figures 1 and 2) RESULTS
Information was available on 138 women aged between 50 and 67 years (mean 57 years). A localisation procedure was necessary in 136 patients because the lesion was impalpable, and in 2 women because the lesion was present in diusely nodular breast tissue. The mammographic ®ndings are presented in Table 1. Ultrasound was performed in 101 cases with a mass seen in 21 of these and no abnormality seen in 80. In 133 patients the localisation was by stereotactic guidance using both a Hawkins I wire and carbon marking. In 4 patients no wire was used and carbon alone injected under stereotactic guidance. In 1 patient the lesion was localised under ultrasound guidance with a Hawkins I needle and carbon injection. In all patients, carbon was accurately placed. The pathologist de®ned accurate placement if carbon was visible microscopically in the same high power ®eld as the lesion. The Hawkins needle was also accurately placed in those patients in whom a wire was also inserted, passing in all cases within 10 mm of the lesion. Histology of the surgically excised abnormalities is presented in Tables 2 and 3. There were 92 malignant and 46 benign lesions giving a malignant to benign ratio of 2 : 1. Preoperative ®ne needle aspiration cytology (FNAC) was obtained in all patients. The relation of preoperative cytology to postoperative histology is shown in Table 4. In all patients, the volume of excised tissue was available. Figure 3 shows the mean volume of tissue removed according to the preoperative cytology grade with the minimum and maximum shown. The mean volume of tissue excised in the C2 (benign) group was 28.1 cm3 (range 19.8± 36.4 cm3) and in the C5 (malignant) group was 179.7 cm3 (range 146±213 cm3). Overall the mean volume of tissue excised was 125.6 cm3. Using the Mann-Whitney two sample test, comparing the volume of excised tissue between the dierent preoperative cytology groups, there was a statistically signi®cant dierence between C2 (benign) and C3 ( probably benign) compared with C4 ( probably malignant) and C5 (malignant) (P 5 0.0001 extremely signi®cant).Statistically signi®cant dierences were also found between C2 and C5 (P 5 0.0001), between C1, C2, C3 and C4 compared with C5 (P 5 0.0001) and also between C4 and C5 (P 5 0.0001). The mean volume of tissue removed in those patients with a ®nal benign pathological diagnosis was 35.1 cm3 and in the patients with a ®nal diagnosis of malignancy was 121.2 cm3, the dierence was extremely statistically signi®cant (P 5 0.0001). In 78 patients the weight of the excised specimen was available. Figure 4 shows the mean weight of tissue removed according to the preoperative cytology results with minimum and maximum also shown. The mean overall weight of tissue removed was 42 g. The mean
THE USE OF CARBON SUSPENSION AS AN ADJUNCT TO WIRE LOCALISATION OF IMPALPABLE BREAST LESIONS
Fig. 1 ± (a) Photograph at operation and (b) operative specimen showing hookwire and carbon at site of lesion.
939
940
CLINICAL RADIOLOGY
Fig. 2 ± A low power magni®cation view demonstrates the carbon track in the centre of an area of ductal carcinoma in site.
weight of tissue excised in the C2 group was 17 g (range 11 g±23 g) and in the C5 group was 70 g (range 50 g±90 g). As with the volume of tissue removed, using the MannWhitney two-sample test there was a statistically signi®cant dierence between the weights of tissue excised, depending on the preoperative cytology when the groups were compared. (C2 and C3 compared with C4 and C5, C2 compared with C5, and C1, C2, C3 and C4 compared with C5 and C4 compared with C5 (P 5 0.0001). The mean weight of tissue removed in those patients with a ®nal benign diagnosis was 21 g and in those with a ®nal malignant diagnosis was 53 g which was a statistically signi®cant dierence (P 5 0.0001).
DISCUSSION
The most frequently used method of marking an impalpable breast lesion to enable surgical excision is by the insertion of a hookwire into the lesion usually under
stereotactic mammographic or ultrasound guidance. The surgeon follows the wire down to the lesion and excises the area surrounding the hook [2±20]. The localising wire should be inserted through the skin as near as possible to the site of planned surgical incision with the most direct route possible to the lesion. Localisation techniques using hookwires are reasonably comfortable for the patient, usually dicult to dislodge and gentle traction on the needle aids surgery. Disadvantages include a signi®cant number of miss rates, the wire may migrate or retract and so surgery must be performed soon after insertion, the wire may be dislodged before or during surgery or it may be accidentally cut. Some types of wire are not easily seen or palpated in the breast tissue at surgery. Other reported complications include pneumothorax, haematomas, wound infections and retained wire fragments post operatively [2,6,17,27]. Table 2 ± Histology of benign lesions Number of patients
Table 1 ± Mammographic ®ndings Mammographic ®ndings
No microcalci®cation Microcalci®cation
Mass Stellate lesion Assymetrical density Microcalci®cation only
39 10 9
19 3 2 56
Fibrocystic disease Fibroadenoma Atypical ductal hyperplasia Lymph node Intraductal papillomatosis Intraductal papilloma Radial scar Normal breast Total
15 4 1 1 1 2 20 2 46
THE USE OF CARBON SUSPENSION AS AN ADJUNCT TO WIRE LOCALISATION OF IMPALPABLE BREAST LESIONS
941
Table 3 ± Histology of malignant lesions No. Number of patients Ductal carcinoma in situ (DCIS) Invasive ductal carcinoma (IDC) Invasive lobular carcinoma (ILC) Invasive tubular carcinoma (ITC) Invasive medullary carcinoma (IMC) IDC and ILC ILC and ITC Total
38 37 7 4 1 4 1 92 Fig. 3 ± Mean volume of tissue excised according to preoperative cytology result. Vertical lines show the minimum and maximum volumes removed in each category.
Table 4 ± Pre and postoperative pathology Preoperative cytology
C1 C2 C3 C4 C5
(n 14) (n 30) (n 24) (n 22) (n 48)
non diagnostic `de®nitely benign' `probably benign' `probably malignant' `de®nitely malignant'
Postoperative histology Benign
Malignant
7 21 13 5 0
7 9 11 17 48
Absolute sensitivity of C5 cytology 52%, Speci®city 100%, Complete sensitivity of C4/C5 cytology 70%, Acellular 10%.
In 1983 Svane described a stereotactic technique for preoperative localisation of breast lesions using carbon suspension [8]. Under stereotactic guidance, 0.3 to 1 ml of an aqueous suspension of carbon powder was injected into the lesion with continued injection whilst withdrawing the needle resulted in a distinct trail of carbon to a small tattoo on the skin. The surgeon then followed the carbon tract to the lesion. Carbon is completely inert. Although carbon granulomas have been reported in the breast following injection for preoperative localisation [30], carbon does not usually cause any pathological reaction in the breast tissue and so pathological interpretation is undisturbed. Carbon does not diuse into the breast tissue unlike methylene blue dye [5,17,27] and unlike wires, cannot be displaced and so the localisation procedure does not need be performed close to the time of surgery. In Svane's original article, 31 of the 56 lesions were removed at other hospitals some time after localisation and one was removed 57 days after localisation with no diusion of the carbon tract [8]. As the lesion site can be marked extremely accurately, it can be identi®ed and excised with minimal tissue excision, particularly important if the lesion is benign [8]. Others have described the use of carbon to localise impalpable breast lesions and have con®rmed its advantages [7,17,18,24,28,29]. A recent paper describes the use of carbon marking at the time of stereotactic core biopsy in 247 lesions with subsequent successful lesion excision in 132 patients. Specimen sizes compared favourably to hookwire localisation, and there were no changes attributable to
Fig. 4 ± Mean weight of tissue excised according to preoperative cytology result. Vertical lines show the minimum and maximum weight removed in each category.
carbon in 68 lesions followed mammographically [29]. Surgeons have found the technique extremely useful as the carbon is easily seen and have readily adopted the carbon method [18]. We describe the use of carbon suspension as an adjunct to guidewire localisation. The method adds very little time or cost to the procedure and is well tolerated by the patient. We use a Hawkins Type 1 needle, which is rigid, and its tip is easily palpated in the breast by movement of its proximal end [19]. We inject a small amount of carbon into the lesion under imaging guidance. We do not leave a tract to the skin unless the lesion is very super®cial and no guidewire is inserted. The carbon at the site of the mammographic abnormality is readily identi®ed at surgery, increasing the surgeon's con®dence that the correct area has been excised. This is particularly useful where the specimen radiograph may be dicult to interpret, for example in subtle distortions. Appropriate margin excision in lesions proven to be malignant prior to operation may also be easier to gauge at surgery because the carbon is visualised. Multiple injections of carbon may be made to mark boundaries of more diuse areas of microcalci®cation increasing the
942
CLINICAL RADIOLOGY
surgical con®dence that all the abnormality has been excised. Most localisation procedures performed in our unit were using stereotactic guidance, including 20 of the 21 patients in whom a mass was demonstrated on ultrasound. This re¯ects our preference for mammographic guidance because of the high percentage of microcalci®cation and small masses requiring excision, which were dicult to demonstrate with con®dence on ultrasound. Ultrasound is however used successfully to localise impalpable breast abnormalities and guide the placement of a wire or simply to mark a spot on the skin overlying a super®cial lesion [21±24]. Ultrasound to guide the injection of carbon suspension has also been described [24]. Failure to remove all or part of a non-palpable, mammographically detected abnormality is reported in up to 17.9% of needle localisation breast biopsy procedures [3]. Others report failure rates of 2.5% 10%. [2,5,15] Factors associated with unsuccessful excision included lesion type (less successful in isolated microcalci®cation), small lesion size, more than one lesion per breast, accuracy of needle placement, small volume of tissue surgically removed, dislodgement of the needle, poor communication between the radiologist and surgeon and type of hook wires used [2,3,15]. To increase the success rate of excision of the mammographic abnormality at the initial surgical procedure, a specimen radiograph is recommended and if this fails to show the abnormality, further excision can be performed [2,5]. Both false positives and false negatives may occur with the use of specimen radiographs [10,11]. Specimen radiography is always performed in our institution. In our study, all of the mammographic abnormalities were accurately localised leading to successful surgical excision in all 138 patients (100% success rate). The carbon was present in all lesions. The wire, where used, was within 10 mm of the lesion, meeting current guidelines [1,31]. Since the introduction of guidelines on specimen weight from the BASO in October 1995 [1] we have routinely measured specimen weight. This data was available in 78 patients. Since specimen weight was not available on patients prior to 1995 we also estimated specimen volumes from the measurements of length, breadth and height on the pathology reports in all 138 patients. The mean volume and weight excised in patients who subsequently had a benign diagnosis was 35.1 cm3 and 21 g, the latter comparing well with the recommended weight of 20 g [1].
The mean volume and weight of tissue excised in those patients who subsequently had a malignant diagnosis was 121.1 cm3 and 53 g respectively. The dierence in both weight and volume of tissue removed between the two groups was extremely statistically signi®cant (P 5 0.0001). With the increased use of preoperative diagnosis by either FNAC or core biopsy, wire localisation is increasingly being used with the aim of performing a primary therapeutic procedure, the surgeon being guided in the extent of resection by preoperative pathology [20]. This is obviously of bene®t in terms of reduced health costs and reduced morbidity due to local complications and psychological concern. In one study of symptomatic and screen detected abnormalities, localisation biopsy was successful in removing the abnormality in all patients and was both diagnostic and therapeutic in 79% of patients with malignancy [20]. It was concluded that for the majority of patients presenting with impalpable lesions detected on mammography, the initial localisation procedure should be both diagnostic and therapeutic. Others have examined factors in¯uencing the therapeutic success of localisation biopsy [3,4]. Therapeutic success, de®ned as the need for no further local surgery was 41% and was related to the accuracy of preoperative needle localisation, preoperative cytology result of C4 or C5 and increased specimen weight [4]. Others report similar ®ndings, with smaller biopsy weight (550 g), larger tumours and the absence of a preoperative diagnosis of malignancy associated with inadequate excision [34]. A preoperative diagnosis of C1-4 was more likely to be associated with inadequate resection margins in those patients subsequently shown to have a malignancy. Residual disease was found in 63% of patients who had reexcision. The assessment of `therapeutic success' is often determined by the presence of tumour at the surgical margins and clearance [4,16,20]. The presence of disease at the resection margin is associated with high recurrence rates. Unfortunately, there is no consensus as to what constitutes an adequate resection margin in the surgery of breast cancer with recommendations ranging from more than 1 mm to 20 mm [1,32,33]. Surgical margins that are clear of tumour do not exclude the presence of residual carcinoma [14,35,36]. Decisions regarding further treatment after a localisation procedure in our unit are made at a multidisciplinary
Table 5 ± Excised specimen margins and further surgery No. patients
Reexcision
Mastectomy
Specimen margins clear of tumour Specimen margins not clear of tumour
70 22*
Lesion not in specimen
0
0 14 10 - no residual malignancy 4 - residual malignancy
1 (no residual malignancy) 7 4 - no residual malignancy 3 - residual malignancy
*Note: 1 patient had an 8 mm well dierentiated Grade 1 tubular carcinoma which was within 1 mm of resection margin. Because of medical problems it was elected not to perform further surgery.
THE USE OF CARBON SUSPENSION AS AN ADJUNCT TO WIRE LOCALISATION OF IMPALPABLE BREAST LESIONS
meeting and are determined by many factors which include the patient and her wishes, mammographic ®ndings and the pathology, including the pathological subtype, degree of invasiveness and the resection margin. We de®ned therapeutic success as the need for no further surgery after multidisciplinary review. We found a high therapeutic success of localisation biopsy in our patient group. Further surgery was required in only 22 of 92 patients with malignancy (20%). This compares very favourably with previous studies which report a re-excision rate between 41% 59% [4,34]. Only one of our patients with clear margins had further surgery, requesting a mastectomy but no residual carcinoma was present. In those 22 patients with margins which were not clear no whom further surgery was performed (Table 5), residual carcinoma was found in 7 (39%). This compares very favourably to other studies in which residual carcinoma was found in 63% of patients [34]. If the preoperative FNAC or core biopsy con®rms that the mammographic abnormality is malignant, the women can be counselled and staged before having a single, therapeutic operation. We demonstrated that as the preoperative cytology was more suspicious of malignancy (C4), with highly suspicious radiology, the surgeon removes a statistically signi®cant larger amount of tissue, both when measured by volume and weight (Figures 3 and 4). In all patients in whom the preoperative cytology was C5, a diagnosis of malignancy was con®rmed at surgery. Of concern was the signi®cant rate of malignancy in those patients with a preoperative diagnosis of C2 (30% malignant) and C3 (46% malignant). This can be partly explained by the learning curve for accurate cytology sampling in the early days of the screening programme and has been addressed in our unit by a move to core biopsy. Benign biopsies are an inevitable part of any screening programme and add signi®cantly to both the ®nancial and psychological cost of screening. The ratio of benign to malignant biopsies should be kept to a minimum and should not exceed 1 : 1 [1,37,38]. In our study the benign to malignant ratio was 1 : 2 with 67% of biopsies malignant. BASO suggests that over 70% of screen detected carcinomas should have a de®nite preoperative diagnosis [1]. The introduction of core biopsy at assessment has resulted in much higher rates of preoperative de®nitive diagnosis and a lower rate of benign surgical biopsy [39±41]. In one unit this resulted in an increase in the preoperative diagnosis rate from 72% with FNAC to 90% with core biopsy and led to a 64% reduction in diagnostic biopsies for screen detected cancer [39]. In our study, in those patients with a ®nal diagnosis of malignancy, there was a de®nite diagnosis of cancer preoperatively in 48 patients (52%) and a probable or de®nite diagnosis of cancer in 65 patients (71%). Our ®gures do not include those patients with malignant lesions who do not require localisation. If this group of patients is included our pre-operative de®nite diagnosis rate is over 70% meeting the BASO guidelines [1]. Recently a more aggressive policy of core biopsy has been introduced
943
and this should improve this rate still further. One of the problems of core biopsy we have found is that all of the radiological abnormality may be excised. This is a particular problem with small clusters of microcalci®cations. Carbon has a role as a marker at the time of core biopsy, allowing subsequent localisation even if the radiological abnormality is removed by the cores. [29] Our method of needle localisation biopsy appears to be an extremely accurate technique for surgical excision of mammographically detected breast abnormalities, adding very little in time or cost to the procedure. The main advantages of this method are of improved accuracy, with smaller biopsies in patients without a pre-operative diagnosis of malignancy, and better cosmesis since the site of the surgical scar is not limited by the site of wire insertion. In addition the surgeon has much greater con®dence that the lesion has been removed, particularly if the specimen radiograph is equivocal. Pathologically, the black carbon is easily seen at the site of mammographic abnormality, staining the abnormal breast tissue and the pathologist ®nds the carbon extremely helpful in guiding him or her to exactly where the mammographic abnormality lies within the specimen. It is essential that there is close communication between radiologist and surgeon. We have also found needle localisation with additional carbon injection to be of therapeutic value in 79% of our patients. The extent of surgical resection is guided by preoperative triple assessment, part of which is preoperative FNAC. We have shown how the extent of surgical resection is related to preoperative cytology. The introduction of preoperative core biopsy in our department we anticipate will increase our preoperative rate of de®nite diagnosis and should be of further use in guiding the extent of resection and should increase our rate of therapeutic localisation procedures further. Acknowledgements. We would like to thank Carol Hay from the Audit Department, Addenbrooke's Hospital, Cambridge for help with data analysis. REFERENCES 1 The Breast Surgeons Group of the British Association of Surgical Oncology. Guidelines for surgeons in the management of symptomatic breast disease in the United Kingdom. Eur J Surg Oncol 1995;21(supplement A): 1±13. 2 Rissanen TJ, Makarainen HP, Mattila SI, Karttunen AI, Kiviniemi HO, Kallioinen MJ, Kaarela OI. Wire localized biopsy of breast lesions: a review of 425 cases found in screening or clinical mammography. Clin Radiol 1993;47:14±22. 3 Jackman RJ, Marzoni FA. Needle-localized breast biopsy: why do we fail?. Radiology 1997;204:677±684. 4 Chadwick DR, Shorthouse AJ. Wire-directed localization biopsy of the breast: an audit of results and analysis of factors in¯uencing therapeutic value in the treatment of breast cancer. Eur J Surg Oncol 1997;23:128±133. 5 Weyant M, Carroccio A, Tartter PI, Hermann G, Levykh R, Robenzadeh A, Adawale A. Determinants of success with spot localization biopsy of the breast. J Am Coll Surg 1995;181:521±524. 6 Querci della Rovere G, Benson JR, Morgan M, Warren R, Patel A. Localization of impalpable breast lesions ± a surgical approach. Eur J Surg Oncol 1996;22:478±482. 7 Canavese G, Catturich A, Vecchio C, Tomei D, Estienne M, Moresco L, Imperiale A, Parodi GC, Massa T, Badellino F.
944
8 9
10
11 12 13
14 15 16
17 18 19 20
21 22 23
24
CLINICAL RADIOLOGY
Pre-operative localization of non-palpable lesions in breast cancer by charcoal suspension. Eur J Surg Oncol 1995;21:47±49. Svane G. A stereotactic technique for preoperative marking of nonpalpable breast lesions. Acta Radiol 1983;24:145±151. Higgins G, Abdulkareen A, Lout® A, Gagnon JH, Shibata HR. Productivity of needle localization to facilitate excision of non palpable, mammographically suspicious lesions. CJS,1991;34: 287±289. Lee CH, Carter D. Detecting residual tumour after excisional biopsy of impalpable breast carcinoma: ecacy of comparing preoperative mammograms with radiographs of the biopsy specimen. Am J Roentgenol 1995;164:81±86. Hasselgren P, Hummel RP, Georgian-Smith D, Fieler M. Breast biopsy with needle localization: accuracy of specimen X-ray and management of missed lesions. Surgery 1993;114:836±842. Stein MA, Karlan M. Immediate postoperative mammogram for failed surgical excision of breast lesions. Radiology 1991;178: 159±162. Hasselgren P, Hummel RP, Fieler M. Breast biopsy with needle localization: in¯uence of age and mammographic feature on the rate of malignancy in 350 nonpalpable breast lesions. Surgery 1991;110:623±628. Acosta JA, Greenlee JA, Gubler KD, Goepfert CJ, Ragland JJ. Surgical margins after needle-localization breast biopsy. Am J Surg 1995;170:643±646. Yankaskas BC, Knelson MH, Abernethy ML, Cuttino JT, Clark RL. Needle localization biopsy of occult lesions of the breast. Experience in 199 cases. Invest Radiol 1988;23:729±733. Hastrich DJ, Dunn JM, Armstrong JS, Davies JD, Davies ZD, Webb AJ, Farndon JR. Diagnostic and therapeutic aspects of ®newire localization biopsy for impalpable breast cancer. Br J Surg 1992;79:1038±1041. Dufrane P, Mazy G, Vanhaudenaerde C. Prebiopsy localization of non palpable breast cancer. JBR-BTR 1990;73:401±404. Langlois S, Carter ML. Carbon localization of impalpable mammographic abnormalities. Australas Radiol 1991;35:237±241. Walker TM, Ross HB. Technical notes. Localization of impalpable breast lesions using the Hawkins needle. Br J Radiol 1992;65: 929±931. Allen MJ, Thompson WD, Stuart RC, Gill PT, Walton EW, Karczenski K, Homan J, Peel ALG. Management of non palpable breast lesions detected mammographically. Br J Surg 1994;81: 543±545. Wilson M, Boggis CRM, Mansel RE, Harland RNL. Non-invasive ultrasound localization of impalpable breast lesions. Clin Radiol 1992;47:337±338. Davies AH, Cowan A, Jones P, Watkins RM, Teasdale C. Ultrasound localization of screen detected impalpable breast tumours. J R Coll Surg Edinb 1994;39:353±354. Fornage BD, Ross MI, Singletary SE, Paulus DD. Localization of impalpable breast masses: value of sonography in the operating room and scanning of excised specimens. Am J Roentgenol 1994;163:569±573. Canavese G, Catturich A, Vecchio C, Tomei D, Estienne M, Massa T, Parodi G, Bruzzone M, Imperiale A, Sardanelli F, Badellino F.
25
26 27 28 29 30 31 32 33 34
35 36
37
38
39
40 41
Non palpable breast lesions: Preoperative sonographic localization. Int Surg 1995;80:283±286. Brenner RJ, Shellock FG, Rothman BJ, Giuliano A. Technical note: magnetic resonance imaging-guided preoperative breast localization using a ``freehand technique''. Br J Radiol 1995;68: 1095±1098. Kuhl CK, Elevelt A, Leutner CC, Gieseke J, Pakos E, Schild HH. Interventional breast MR imaging: clinical use of a stereotactic localization and biopsy device. Radiology 1997;204:667±675. Rappaport W, Thompson S, Wong R, Leong S, Villar H. Complications associated with needle localization biopsy of the breast. Surg Gynecol and Obst 1991;172:303±306. Potchen EJ, Sierra A, Mackenzie C, Osuch J. Svane localization of non-palpable breast lesions. Lancet 1991;338:816. Mullen DJ, Eisen RN, Newman RD, Perrone PM, Wilsey JC. The Use of Carbon Marking After Stereotactic Large Core Needle Breast Biopsy. Radiology 2001;218:255±260. Patrikeos A, Wylie EJ, Bourke A, Frost F. Imaging of carbon granulomas of the breast following carbon tract localization. Clin Radiol 1998;53:854±848. Board of Faculty of Clinical Radiology, The Royal College of RadiologistsGuidance on Screening and Symptomatic Breast Imaging. RCR Ref No BFCR(99) 12. Quality Assurance Guidelines for Surgeons in Breast Cancer Screening. NHSBSP Publication No 20 1992 revised 1996. The Nottingham Prognostic Index in Primary Breast Cancer. Breast Cancer Research and Treatment 1992;22:207±219. Chinyama CN, Davies JD, Rayter Z, Farndon JR. Factors aecting surgical margin clearance in screen-detected breast cancer and the eect of cavity biopsies on residual disease. Eur J Surg Oncol 1997;23:123±127. McKinney CD, Frierson HF, Fechner RE, Wilhelm MC, Edge SB. Pathologic ®ndings in non palpable invasive breast cancer. Am J Surg Path 1992;16:33±36. Schnitt SJ, Connolly JL, Khettry U, Mazoujian G, Brenner M, Silver B, Recht A, Beadle G, Harris JR. Pathologic ®ndings on reexcision of the primary site in breast cancer patients considered for treatment by primary radiation therapy. Cancer 1987;59: 675±681. Spencer NJB, Evans AJ, Galea M, Sibbering DM, Yeoman LJ, Pinder SE, Ellis IO, Elston CW, Blamey RW, Robertson JFR, Wilson ARM. Pathological-radiological correlations in benign lesions excised during a breast screening programme. Clin Radiol 1994;49:853±856. Burnett SJD, Ng YY, Perry NM, Gilmore OJA, Allum WH, Carpenter R, Wells CA. Benign biopsies in the prevalent round of breast screening: a review of 137 cases. Clin Radiol 1995;50: 254±258. Litherland JC, Evans AJ, Wilson ARM, Kollias J, Pinder SE, Elston CW, Ellis IO, Yeoman LJ. The impact of core biopsy on preoperative diagnosis rate of screen detected breast cancers. Clin Radiol 1996;51:562±565. Burns RP. Image guided breast biopsy. Am J Surg 1997;173:9±11. Parker SH, Burbank F. A practical approach to minimally invasive breast biopsy. Radiology 1996;200:11±20.
The Use of Carbon Suspension as an Adjunct to Wire Localisation of Impalpable Breast Lesions H . A . M O S S , S . J . B A R T E R , M . NAYA G A M , D. L AW R EN C E , M . P I T TA M *Department of Radiology, Addenbrooke's Hospital, Hills Road, Cambridge, {Departments of Radiology, {Pathology and }Surgery, Beds and Herts Breast Screening Service, Luton and Dunstable Hospital NHS Trust, Lewsey Road, Luton Bedfordshire LU4 0DZ, UK Received: 12 October 2001 Revised: 27 March 2002
Accepted: 5 April 2002
AIM: To determine the accuracy and therapeutic success of localisation of impalpable breast lesions by hookwire with additional lesion marking with carbon suspension to mark screen detected abnormalities requiring surgical excision. MATERIALS AND METHODS: Retrospective review of all breast localisation procedures performed in our unit on women with a screen detected abnormality requiring excision over a 7 year period. RESULTS: One hundred and thirty eight women underwent breast localisation procedures. All of the mammographic abnormalities were excised at the initial surgical procedure. The benign to malignant ratio was 1 : 2. Pre-operative cytology was used to guide the extent of surgical excision, with clear margins in 70 of the 92 patients (75%) with malignancy. Twenty patients had further surgery: mastectomy in 7 and further local excision in 14. The localisation procedure was a therapeutic success in the local excision of malignancy in 73 of the 92 patients (79%) with malignancy. CONCLUSION: This method of localisation biopsy is an accurate technique for surgical excision of mammographically detected impalpable abnormalities. The surgeon is able to choose the site of surgical incision to give the best cosmetic result, the lesion is easier to identify at operation and the con®dence that the abnormality has been excised is improved. Moss H. A. et al. (2002) Clinical Radiology 57, 937±944. # 2002 The Royal College of Radiologists Published by Elsevier Science Ltd. All rights reserved. Key words: carbon, breast biopsy, breast surgery, breast neoplasms, localisation.
INTRODUCTION
The National Health Service Breast Screening Programme has resulted in the detection of numerous impalpable lesions in asymptomatic women. After assessment clinically, with imaging and ®ne needle aspiration cytology (FNAC) or core biopsy, some of these lesions require surgical excision biopsy. Accurate preoperative localisation is essential to enable surgical removal of the area of concern whilst also achieving an optimal cosmetic result and minimum patient morbidity. Guidelines from the British Association of Surgical Oncologists (BASO) advise a maximum biopsy weight of 20 g for lesions not proven to be malignant preoperatively [1]. A variety of localisation techniques have been described using mammographic stereotactic [2±20], ultrasound [7,17,19,21±24] and magnetic resonance imaging guidance [25,26]. Methods of marking the abnormality to guide the surgeon to the lesion include insertion of a hookwire Address for correspondence and guarantor of study: Dr SJ Barter Beds and Herts Breast Screening Service, Luton and Dunstable Hospital NHS Trust, Lewsey Road, Luton, Bedfordshire, LU4 ODZ, UK. Fax: 01234-792106; E-mail: [email protected] 0009-9260/02/$35
[2±20,22], injection of a dye such a methylene blue [5,7,17,27] or carbon [7,8,17,18,24,28,29] or marking the skin [17,21,22]. The most frequently used technique is hookwire insertion into the lesion under imaging control, using stereotactic mammographic guidance [2±20]. The patient is then taken to the operating theatre where the surgeon commonly makes an incision at the site of hookwire insertion and follows the wire down to the hook, which is excised along with the lesion. Problems encountered with this technique include signi®cant rates of failure to excise the lesion in up to 17.9% of patients. Studies have reported success rates of localisation ranging from 82.1% to 93.7% [2,3,5,15,27]. In addition the surgeon must make the incision close to the point of insertion of the wire which may not give the best cosmetic result, and which may result in a long track to the abnormality. Also, specimen radiographs may be dicult to interpret with small and subtle lesions of the kind that are detected by breast screening. In 1993, Svane described the use of carbon suspension for the localisation of impalpable mammographic abnormalities [8]. A 4% aqueous suspension of carbon was injected into the lesion under stereotactic guidance. As the needle was
# 2002 The Royal College of Radiologists Published by Elsevier Science Ltd. All rights reserved.
938
CLINICAL RADIOLOGY
withdrawn, carbon was continuously injected to produce a tract of carbon from the lesion to the skin, where the carbon produced a small mark or tattoo. The surgeon was then able to follow the carbon track to the lesion and excise it. Carbon was found to be ideal for marking since it did not diuse into surrounding tissue and its inert properties meant that the localisation procedure did not have to be carried out immediately prior to surgery. In our unit, we use a combination of a rigid hookwire (Hawkins Type I) and carbon suspension to localise impalpable lesions. We describe this technique, its advantages and the diagnostic and therapeutic results of breast localisation using this method. MATERIALS AND METHOD
We retrospectively reviewed all breast localisation procedures performed on women in the screening population who attended our Breast Screening Service for localisation of screen detected impalpable lesions between October 1991 and December 1998. The clinical and imaging ®ndings, pre-operative cytology, and method of localisation was collected from the Breast Screening database. Histology of the excised specimen, its size, weight and excision margins were obtained from the pathology reports. Any further surgical procedures such as re-excision or mastectomy were noted. Statistical analysis was performed using the MannWhitney two sample tests. A combination of carbon injection and the hookwire technique was used to localise the mammographic abnormality, either using stereotactic control (Siemens Mammomat) or ultrasound guidance (Siemens Sonoline). The skin was anaesthetised with 1% lignocaine and a 21 gauge, 80 mm needle (Guilford needle, Mediplus Ltd) was inserted into the lesion and its position checked with swing views. 0.2±0.4 mls of sterile carbon suspension 40 mg/mls (Apoteket Laboratories) was then injected into the lesion. The needle was then withdrawn and a Hawkin Type I hookwire inserted into the lesion without moving the patient. The position of the hookwire was checked by two mammographic views in orthogonal planes and a diagram made of the needle position in the patient's notes. In 6 patients with segmental microcalci®cation several injections of carbon were made to de®ne the posterior and lateral margins of the mammographic abnormality. If the mammographic abnormality was super®cial, no hookwire was inserted but a carbon track was made from the lesion to the skin surface by constantly injecting a small amount of carbon suspension as the needle was withdrawn, leaving a tattoo on the skin. Surgical excision was performed under general anaesthetic on the same day. As the Hawkins wire is rigid the surgeon is able to palpate its tip and choose the incision site for best approach to the lesion and cosmesis. Using gentle traction on the hookwire, the surgeon dissects down to the tip to locate the carbon. The tissue containing the carbon is then excised and a specimen radiograph taken using mild compression. The result of the specimen radiograph is communicated to the surgeon by the radiologist who performed the localisation procedure to discuss if excision is
adequate. The histology results are reviewed at a weekly multidisciplinary meeting and further management discussed. (Figures 1 and 2) RESULTS
Information was available on 138 women aged between 50 and 67 years (mean 57 years). A localisation procedure was necessary in 136 patients because the lesion was impalpable, and in 2 women because the lesion was present in diusely nodular breast tissue. The mammographic ®ndings are presented in Table 1. Ultrasound was performed in 101 cases with a mass seen in 21 of these and no abnormality seen in 80. In 133 patients the localisation was by stereotactic guidance using both a Hawkins I wire and carbon marking. In 4 patients no wire was used and carbon alone injected under stereotactic guidance. In 1 patient the lesion was localised under ultrasound guidance with a Hawkins I needle and carbon injection. In all patients, carbon was accurately placed. The pathologist de®ned accurate placement if carbon was visible microscopically in the same high power ®eld as the lesion. The Hawkins needle was also accurately placed in those patients in whom a wire was also inserted, passing in all cases within 10 mm of the lesion. Histology of the surgically excised abnormalities is presented in Tables 2 and 3. There were 92 malignant and 46 benign lesions giving a malignant to benign ratio of 2 : 1. Preoperative ®ne needle aspiration cytology (FNAC) was obtained in all patients. The relation of preoperative cytology to postoperative histology is shown in Table 4. In all patients, the volume of excised tissue was available. Figure 3 shows the mean volume of tissue removed according to the preoperative cytology grade with the minimum and maximum shown. The mean volume of tissue excised in the C2 (benign) group was 28.1 cm3 (range 19.8± 36.4 cm3) and in the C5 (malignant) group was 179.7 cm3 (range 146±213 cm3). Overall the mean volume of tissue excised was 125.6 cm3. Using the Mann-Whitney two sample test, comparing the volume of excised tissue between the dierent preoperative cytology groups, there was a statistically signi®cant dierence between C2 (benign) and C3 ( probably benign) compared with C4 ( probably malignant) and C5 (malignant) (P 5 0.0001 extremely signi®cant).Statistically signi®cant dierences were also found between C2 and C5 (P 5 0.0001), between C1, C2, C3 and C4 compared with C5 (P 5 0.0001) and also between C4 and C5 (P 5 0.0001). The mean volume of tissue removed in those patients with a ®nal benign pathological diagnosis was 35.1 cm3 and in the patients with a ®nal diagnosis of malignancy was 121.2 cm3, the dierence was extremely statistically signi®cant (P 5 0.0001). In 78 patients the weight of the excised specimen was available. Figure 4 shows the mean weight of tissue removed according to the preoperative cytology results with minimum and maximum also shown. The mean overall weight of tissue removed was 42 g. The mean
THE USE OF CARBON SUSPENSION AS AN ADJUNCT TO WIRE LOCALISATION OF IMPALPABLE BREAST LESIONS
Fig. 1 ± (a) Photograph at operation and (b) operative specimen showing hookwire and carbon at site of lesion.
939
940
CLINICAL RADIOLOGY
Fig. 2 ± A low power magni®cation view demonstrates the carbon track in the centre of an area of ductal carcinoma in site.
weight of tissue excised in the C2 group was 17 g (range 11 g±23 g) and in the C5 group was 70 g (range 50 g±90 g). As with the volume of tissue removed, using the MannWhitney two-sample test there was a statistically signi®cant dierence between the weights of tissue excised, depending on the preoperative cytology when the groups were compared. (C2 and C3 compared with C4 and C5, C2 compared with C5, and C1, C2, C3 and C4 compared with C5 and C4 compared with C5 (P 5 0.0001). The mean weight of tissue removed in those patients with a ®nal benign diagnosis was 21 g and in those with a ®nal malignant diagnosis was 53 g which was a statistically signi®cant dierence (P 5 0.0001).
DISCUSSION
The most frequently used method of marking an impalpable breast lesion to enable surgical excision is by the insertion of a hookwire into the lesion usually under
stereotactic mammographic or ultrasound guidance. The surgeon follows the wire down to the lesion and excises the area surrounding the hook [2±20]. The localising wire should be inserted through the skin as near as possible to the site of planned surgical incision with the most direct route possible to the lesion. Localisation techniques using hookwires are reasonably comfortable for the patient, usually dicult to dislodge and gentle traction on the needle aids surgery. Disadvantages include a signi®cant number of miss rates, the wire may migrate or retract and so surgery must be performed soon after insertion, the wire may be dislodged before or during surgery or it may be accidentally cut. Some types of wire are not easily seen or palpated in the breast tissue at surgery. Other reported complications include pneumothorax, haematomas, wound infections and retained wire fragments post operatively [2,6,17,27]. Table 2 ± Histology of benign lesions Number of patients
Table 1 ± Mammographic ®ndings Mammographic ®ndings
No microcalci®cation Microcalci®cation
Mass Stellate lesion Assymetrical density Microcalci®cation only
39 10 9
19 3 2 56
Fibrocystic disease Fibroadenoma Atypical ductal hyperplasia Lymph node Intraductal papillomatosis Intraductal papilloma Radial scar Normal breast Total
15 4 1 1 1 2 20 2 46
THE USE OF CARBON SUSPENSION AS AN ADJUNCT TO WIRE LOCALISATION OF IMPALPABLE BREAST LESIONS
941
Table 3 ± Histology of malignant lesions No. Number of patients Ductal carcinoma in situ (DCIS) Invasive ductal carcinoma (IDC) Invasive lobular carcinoma (ILC) Invasive tubular carcinoma (ITC) Invasive medullary carcinoma (IMC) IDC and ILC ILC and ITC Total
38 37 7 4 1 4 1 92 Fig. 3 ± Mean volume of tissue excised according to preoperative cytology result. Vertical lines show the minimum and maximum volumes removed in each category.
Table 4 ± Pre and postoperative pathology Preoperative cytology
C1 C2 C3 C4 C5
(n 14) (n 30) (n 24) (n 22) (n 48)
non diagnostic `de®nitely benign' `probably benign' `probably malignant' `de®nitely malignant'
Postoperative histology Benign
Malignant
7 21 13 5 0
7 9 11 17 48
Absolute sensitivity of C5 cytology 52%, Speci®city 100%, Complete sensitivity of C4/C5 cytology 70%, Acellular 10%.
In 1983 Svane described a stereotactic technique for preoperative localisation of breast lesions using carbon suspension [8]. Under stereotactic guidance, 0.3 to 1 ml of an aqueous suspension of carbon powder was injected into the lesion with continued injection whilst withdrawing the needle resulted in a distinct trail of carbon to a small tattoo on the skin. The surgeon then followed the carbon tract to the lesion. Carbon is completely inert. Although carbon granulomas have been reported in the breast following injection for preoperative localisation [30], carbon does not usually cause any pathological reaction in the breast tissue and so pathological interpretation is undisturbed. Carbon does not diuse into the breast tissue unlike methylene blue dye [5,17,27] and unlike wires, cannot be displaced and so the localisation procedure does not need be performed close to the time of surgery. In Svane's original article, 31 of the 56 lesions were removed at other hospitals some time after localisation and one was removed 57 days after localisation with no diusion of the carbon tract [8]. As the lesion site can be marked extremely accurately, it can be identi®ed and excised with minimal tissue excision, particularly important if the lesion is benign [8]. Others have described the use of carbon to localise impalpable breast lesions and have con®rmed its advantages [7,17,18,24,28,29]. A recent paper describes the use of carbon marking at the time of stereotactic core biopsy in 247 lesions with subsequent successful lesion excision in 132 patients. Specimen sizes compared favourably to hookwire localisation, and there were no changes attributable to
Fig. 4 ± Mean weight of tissue excised according to preoperative cytology result. Vertical lines show the minimum and maximum weight removed in each category.
carbon in 68 lesions followed mammographically [29]. Surgeons have found the technique extremely useful as the carbon is easily seen and have readily adopted the carbon method [18]. We describe the use of carbon suspension as an adjunct to guidewire localisation. The method adds very little time or cost to the procedure and is well tolerated by the patient. We use a Hawkins Type 1 needle, which is rigid, and its tip is easily palpated in the breast by movement of its proximal end [19]. We inject a small amount of carbon into the lesion under imaging guidance. We do not leave a tract to the skin unless the lesion is very super®cial and no guidewire is inserted. The carbon at the site of the mammographic abnormality is readily identi®ed at surgery, increasing the surgeon's con®dence that the correct area has been excised. This is particularly useful where the specimen radiograph may be dicult to interpret, for example in subtle distortions. Appropriate margin excision in lesions proven to be malignant prior to operation may also be easier to gauge at surgery because the carbon is visualised. Multiple injections of carbon may be made to mark boundaries of more diuse areas of microcalci®cation increasing the
942
CLINICAL RADIOLOGY
surgical con®dence that all the abnormality has been excised. Most localisation procedures performed in our unit were using stereotactic guidance, including 20 of the 21 patients in whom a mass was demonstrated on ultrasound. This re¯ects our preference for mammographic guidance because of the high percentage of microcalci®cation and small masses requiring excision, which were dicult to demonstrate with con®dence on ultrasound. Ultrasound is however used successfully to localise impalpable breast abnormalities and guide the placement of a wire or simply to mark a spot on the skin overlying a super®cial lesion [21±24]. Ultrasound to guide the injection of carbon suspension has also been described [24]. Failure to remove all or part of a non-palpable, mammographically detected abnormality is reported in up to 17.9% of needle localisation breast biopsy procedures [3]. Others report failure rates of 2.5% 10%. [2,5,15] Factors associated with unsuccessful excision included lesion type (less successful in isolated microcalci®cation), small lesion size, more than one lesion per breast, accuracy of needle placement, small volume of tissue surgically removed, dislodgement of the needle, poor communication between the radiologist and surgeon and type of hook wires used [2,3,15]. To increase the success rate of excision of the mammographic abnormality at the initial surgical procedure, a specimen radiograph is recommended and if this fails to show the abnormality, further excision can be performed [2,5]. Both false positives and false negatives may occur with the use of specimen radiographs [10,11]. Specimen radiography is always performed in our institution. In our study, all of the mammographic abnormalities were accurately localised leading to successful surgical excision in all 138 patients (100% success rate). The carbon was present in all lesions. The wire, where used, was within 10 mm of the lesion, meeting current guidelines [1,31]. Since the introduction of guidelines on specimen weight from the BASO in October 1995 [1] we have routinely measured specimen weight. This data was available in 78 patients. Since specimen weight was not available on patients prior to 1995 we also estimated specimen volumes from the measurements of length, breadth and height on the pathology reports in all 138 patients. The mean volume and weight excised in patients who subsequently had a benign diagnosis was 35.1 cm3 and 21 g, the latter comparing well with the recommended weight of 20 g [1].
The mean volume and weight of tissue excised in those patients who subsequently had a malignant diagnosis was 121.1 cm3 and 53 g respectively. The dierence in both weight and volume of tissue removed between the two groups was extremely statistically signi®cant (P 5 0.0001). With the increased use of preoperative diagnosis by either FNAC or core biopsy, wire localisation is increasingly being used with the aim of performing a primary therapeutic procedure, the surgeon being guided in the extent of resection by preoperative pathology [20]. This is obviously of bene®t in terms of reduced health costs and reduced morbidity due to local complications and psychological concern. In one study of symptomatic and screen detected abnormalities, localisation biopsy was successful in removing the abnormality in all patients and was both diagnostic and therapeutic in 79% of patients with malignancy [20]. It was concluded that for the majority of patients presenting with impalpable lesions detected on mammography, the initial localisation procedure should be both diagnostic and therapeutic. Others have examined factors in¯uencing the therapeutic success of localisation biopsy [3,4]. Therapeutic success, de®ned as the need for no further local surgery was 41% and was related to the accuracy of preoperative needle localisation, preoperative cytology result of C4 or C5 and increased specimen weight [4]. Others report similar ®ndings, with smaller biopsy weight (550 g), larger tumours and the absence of a preoperative diagnosis of malignancy associated with inadequate excision [34]. A preoperative diagnosis of C1-4 was more likely to be associated with inadequate resection margins in those patients subsequently shown to have a malignancy. Residual disease was found in 63% of patients who had reexcision. The assessment of `therapeutic success' is often determined by the presence of tumour at the surgical margins and clearance [4,16,20]. The presence of disease at the resection margin is associated with high recurrence rates. Unfortunately, there is no consensus as to what constitutes an adequate resection margin in the surgery of breast cancer with recommendations ranging from more than 1 mm to 20 mm [1,32,33]. Surgical margins that are clear of tumour do not exclude the presence of residual carcinoma [14,35,36]. Decisions regarding further treatment after a localisation procedure in our unit are made at a multidisciplinary
Table 5 ± Excised specimen margins and further surgery No. patients
Reexcision
Mastectomy
Specimen margins clear of tumour Specimen margins not clear of tumour
70 22*
Lesion not in specimen
0
0 14 10 - no residual malignancy 4 - residual malignancy
1 (no residual malignancy) 7 4 - no residual malignancy 3 - residual malignancy
*Note: 1 patient had an 8 mm well dierentiated Grade 1 tubular carcinoma which was within 1 mm of resection margin. Because of medical problems it was elected not to perform further surgery.
THE USE OF CARBON SUSPENSION AS AN ADJUNCT TO WIRE LOCALISATION OF IMPALPABLE BREAST LESIONS
meeting and are determined by many factors which include the patient and her wishes, mammographic ®ndings and the pathology, including the pathological subtype, degree of invasiveness and the resection margin. We de®ned therapeutic success as the need for no further surgery after multidisciplinary review. We found a high therapeutic success of localisation biopsy in our patient group. Further surgery was required in only 22 of 92 patients with malignancy (20%). This compares very favourably with previous studies which report a re-excision rate between 41% 59% [4,34]. Only one of our patients with clear margins had further surgery, requesting a mastectomy but no residual carcinoma was present. In those 22 patients with margins which were not clear no whom further surgery was performed (Table 5), residual carcinoma was found in 7 (39%). This compares very favourably to other studies in which residual carcinoma was found in 63% of patients [34]. If the preoperative FNAC or core biopsy con®rms that the mammographic abnormality is malignant, the women can be counselled and staged before having a single, therapeutic operation. We demonstrated that as the preoperative cytology was more suspicious of malignancy (C4), with highly suspicious radiology, the surgeon removes a statistically signi®cant larger amount of tissue, both when measured by volume and weight (Figures 3 and 4). In all patients in whom the preoperative cytology was C5, a diagnosis of malignancy was con®rmed at surgery. Of concern was the signi®cant rate of malignancy in those patients with a preoperative diagnosis of C2 (30% malignant) and C3 (46% malignant). This can be partly explained by the learning curve for accurate cytology sampling in the early days of the screening programme and has been addressed in our unit by a move to core biopsy. Benign biopsies are an inevitable part of any screening programme and add signi®cantly to both the ®nancial and psychological cost of screening. The ratio of benign to malignant biopsies should be kept to a minimum and should not exceed 1 : 1 [1,37,38]. In our study the benign to malignant ratio was 1 : 2 with 67% of biopsies malignant. BASO suggests that over 70% of screen detected carcinomas should have a de®nite preoperative diagnosis [1]. The introduction of core biopsy at assessment has resulted in much higher rates of preoperative de®nitive diagnosis and a lower rate of benign surgical biopsy [39±41]. In one unit this resulted in an increase in the preoperative diagnosis rate from 72% with FNAC to 90% with core biopsy and led to a 64% reduction in diagnostic biopsies for screen detected cancer [39]. In our study, in those patients with a ®nal diagnosis of malignancy, there was a de®nite diagnosis of cancer preoperatively in 48 patients (52%) and a probable or de®nite diagnosis of cancer in 65 patients (71%). Our ®gures do not include those patients with malignant lesions who do not require localisation. If this group of patients is included our pre-operative de®nite diagnosis rate is over 70% meeting the BASO guidelines [1]. Recently a more aggressive policy of core biopsy has been introduced
943
and this should improve this rate still further. One of the problems of core biopsy we have found is that all of the radiological abnormality may be excised. This is a particular problem with small clusters of microcalci®cations. Carbon has a role as a marker at the time of core biopsy, allowing subsequent localisation even if the radiological abnormality is removed by the cores. [29] Our method of needle localisation biopsy appears to be an extremely accurate technique for surgical excision of mammographically detected breast abnormalities, adding very little in time or cost to the procedure. The main advantages of this method are of improved accuracy, with smaller biopsies in patients without a pre-operative diagnosis of malignancy, and better cosmesis since the site of the surgical scar is not limited by the site of wire insertion. In addition the surgeon has much greater con®dence that the lesion has been removed, particularly if the specimen radiograph is equivocal. Pathologically, the black carbon is easily seen at the site of mammographic abnormality, staining the abnormal breast tissue and the pathologist ®nds the carbon extremely helpful in guiding him or her to exactly where the mammographic abnormality lies within the specimen. It is essential that there is close communication between radiologist and surgeon. We have also found needle localisation with additional carbon injection to be of therapeutic value in 79% of our patients. The extent of surgical resection is guided by preoperative triple assessment, part of which is preoperative FNAC. We have shown how the extent of surgical resection is related to preoperative cytology. The introduction of preoperative core biopsy in our department we anticipate will increase our preoperative rate of de®nite diagnosis and should be of further use in guiding the extent of resection and should increase our rate of therapeutic localisation procedures further. Acknowledgements. We would like to thank Carol Hay from the Audit Department, Addenbrooke's Hospital, Cambridge for help with data analysis. REFERENCES 1 The Breast Surgeons Group of the British Association of Surgical Oncology. Guidelines for surgeons in the management of symptomatic breast disease in the United Kingdom. Eur J Surg Oncol 1995;21(supplement A): 1±13. 2 Rissanen TJ, Makarainen HP, Mattila SI, Karttunen AI, Kiviniemi HO, Kallioinen MJ, Kaarela OI. Wire localized biopsy of breast lesions: a review of 425 cases found in screening or clinical mammography. Clin Radiol 1993;47:14±22. 3 Jackman RJ, Marzoni FA. Needle-localized breast biopsy: why do we fail?. Radiology 1997;204:677±684. 4 Chadwick DR, Shorthouse AJ. Wire-directed localization biopsy of the breast: an audit of results and analysis of factors in¯uencing therapeutic value in the treatment of breast cancer. Eur J Surg Oncol 1997;23:128±133. 5 Weyant M, Carroccio A, Tartter PI, Hermann G, Levykh R, Robenzadeh A, Adawale A. Determinants of success with spot localization biopsy of the breast. J Am Coll Surg 1995;181:521±524. 6 Querci della Rovere G, Benson JR, Morgan M, Warren R, Patel A. Localization of impalpable breast lesions ± a surgical approach. Eur J Surg Oncol 1996;22:478±482. 7 Canavese G, Catturich A, Vecchio C, Tomei D, Estienne M, Moresco L, Imperiale A, Parodi GC, Massa T, Badellino F.
944
8 9
10
11 12 13
14 15 16
17 18 19 20
21 22 23
24
CLINICAL RADIOLOGY
Pre-operative localization of non-palpable lesions in breast cancer by charcoal suspension. Eur J Surg Oncol 1995;21:47±49. Svane G. A stereotactic technique for preoperative marking of nonpalpable breast lesions. Acta Radiol 1983;24:145±151. Higgins G, Abdulkareen A, Lout® A, Gagnon JH, Shibata HR. Productivity of needle localization to facilitate excision of non palpable, mammographically suspicious lesions. CJS,1991;34: 287±289. Lee CH, Carter D. Detecting residual tumour after excisional biopsy of impalpable breast carcinoma: ecacy of comparing preoperative mammograms with radiographs of the biopsy specimen. Am J Roentgenol 1995;164:81±86. Hasselgren P, Hummel RP, Georgian-Smith D, Fieler M. Breast biopsy with needle localization: accuracy of specimen X-ray and management of missed lesions. Surgery 1993;114:836±842. Stein MA, Karlan M. Immediate postoperative mammogram for failed surgical excision of breast lesions. Radiology 1991;178: 159±162. Hasselgren P, Hummel RP, Fieler M. Breast biopsy with needle localization: in¯uence of age and mammographic feature on the rate of malignancy in 350 nonpalpable breast lesions. Surgery 1991;110:623±628. Acosta JA, Greenlee JA, Gubler KD, Goepfert CJ, Ragland JJ. Surgical margins after needle-localization breast biopsy. Am J Surg 1995;170:643±646. Yankaskas BC, Knelson MH, Abernethy ML, Cuttino JT, Clark RL. Needle localization biopsy of occult lesions of the breast. Experience in 199 cases. Invest Radiol 1988;23:729±733. Hastrich DJ, Dunn JM, Armstrong JS, Davies JD, Davies ZD, Webb AJ, Farndon JR. Diagnostic and therapeutic aspects of ®newire localization biopsy for impalpable breast cancer. Br J Surg 1992;79:1038±1041. Dufrane P, Mazy G, Vanhaudenaerde C. Prebiopsy localization of non palpable breast cancer. JBR-BTR 1990;73:401±404. Langlois S, Carter ML. Carbon localization of impalpable mammographic abnormalities. Australas Radiol 1991;35:237±241. Walker TM, Ross HB. Technical notes. Localization of impalpable breast lesions using the Hawkins needle. Br J Radiol 1992;65: 929±931. Allen MJ, Thompson WD, Stuart RC, Gill PT, Walton EW, Karczenski K, Homan J, Peel ALG. Management of non palpable breast lesions detected mammographically. Br J Surg 1994;81: 543±545. Wilson M, Boggis CRM, Mansel RE, Harland RNL. Non-invasive ultrasound localization of impalpable breast lesions. Clin Radiol 1992;47:337±338. Davies AH, Cowan A, Jones P, Watkins RM, Teasdale C. Ultrasound localization of screen detected impalpable breast tumours. J R Coll Surg Edinb 1994;39:353±354. Fornage BD, Ross MI, Singletary SE, Paulus DD. Localization of impalpable breast masses: value of sonography in the operating room and scanning of excised specimens. Am J Roentgenol 1994;163:569±573. Canavese G, Catturich A, Vecchio C, Tomei D, Estienne M, Massa T, Parodi G, Bruzzone M, Imperiale A, Sardanelli F, Badellino F.
25
26 27 28 29 30 31 32 33 34
35 36
37
38
39
40 41
Non palpable breast lesions: Preoperative sonographic localization. Int Surg 1995;80:283±286. Brenner RJ, Shellock FG, Rothman BJ, Giuliano A. Technical note: magnetic resonance imaging-guided preoperative breast localization using a ``freehand technique''. Br J Radiol 1995;68: 1095±1098. Kuhl CK, Elevelt A, Leutner CC, Gieseke J, Pakos E, Schild HH. Interventional breast MR imaging: clinical use of a stereotactic localization and biopsy device. Radiology 1997;204:667±675. Rappaport W, Thompson S, Wong R, Leong S, Villar H. Complications associated with needle localization biopsy of the breast. Surg Gynecol and Obst 1991;172:303±306. Potchen EJ, Sierra A, Mackenzie C, Osuch J. Svane localization of non-palpable breast lesions. Lancet 1991;338:816. Mullen DJ, Eisen RN, Newman RD, Perrone PM, Wilsey JC. The Use of Carbon Marking After Stereotactic Large Core Needle Breast Biopsy. Radiology 2001;218:255±260. Patrikeos A, Wylie EJ, Bourke A, Frost F. Imaging of carbon granulomas of the breast following carbon tract localization. Clin Radiol 1998;53:854±848. Board of Faculty of Clinical Radiology, The Royal College of RadiologistsGuidance on Screening and Symptomatic Breast Imaging. RCR Ref No BFCR(99) 12. Quality Assurance Guidelines for Surgeons in Breast Cancer Screening. NHSBSP Publication No 20 1992 revised 1996. The Nottingham Prognostic Index in Primary Breast Cancer. Breast Cancer Research and Treatment 1992;22:207±219. Chinyama CN, Davies JD, Rayter Z, Farndon JR. Factors aecting surgical margin clearance in screen-detected breast cancer and the eect of cavity biopsies on residual disease. Eur J Surg Oncol 1997;23:123±127. McKinney CD, Frierson HF, Fechner RE, Wilhelm MC, Edge SB. Pathologic ®ndings in non palpable invasive breast cancer. Am J Surg Path 1992;16:33±36. Schnitt SJ, Connolly JL, Khettry U, Mazoujian G, Brenner M, Silver B, Recht A, Beadle G, Harris JR. Pathologic ®ndings on reexcision of the primary site in breast cancer patients considered for treatment by primary radiation therapy. Cancer 1987;59: 675±681. Spencer NJB, Evans AJ, Galea M, Sibbering DM, Yeoman LJ, Pinder SE, Ellis IO, Elston CW, Blamey RW, Robertson JFR, Wilson ARM. Pathological-radiological correlations in benign lesions excised during a breast screening programme. Clin Radiol 1994;49:853±856. Burnett SJD, Ng YY, Perry NM, Gilmore OJA, Allum WH, Carpenter R, Wells CA. Benign biopsies in the prevalent round of breast screening: a review of 137 cases. Clin Radiol 1995;50: 254±258. Litherland JC, Evans AJ, Wilson ARM, Kollias J, Pinder SE, Elston CW, Ellis IO, Yeoman LJ. The impact of core biopsy on preoperative diagnosis rate of screen detected breast cancers. Clin Radiol 1996;51:562±565. Burns RP. Image guided breast biopsy. Am J Surg 1997;173:9±11. Parker SH, Burbank F. A practical approach to minimally invasive breast biopsy. Radiology 1996;200:11±20.