Radioguided occult lesion localisation for wide local excision, excision biopsies and in combination with radioisotope sentinel lymph node localisation (SNOLL) – 10 year experience of a single centre

Abstract

Background

Various techniques are used to localise non-palpable breast cancer and identify sentinel lymph nodes (SLN). Seed-based localisations can be expensive, may dislodge and require special intraoperative equipment. We report our experience of using radioguided occult lesion localisation (ROLL) alone and with SLN (SNOLL)

Methods

This is a retrospective review of all patients undergoing breast excision procedures using ROLL between January 2008 and 2018. On the morning of surgery or the day before, patients had 10 MBq Technetium 99 m colloid injected into the centre of the breast lesion under ultrasound guidance with a further 10 MBq injected into the periareolar skin if SLN was planned. Surgery was performed with the aid of a gamma probe to detect the radioisotope signal in the breast and axilla and specimen X-rays of the breast excision specimen were performed with cavity shaves considered if appropriate.

Results

1073 ROLL-guided excisions were performed in 1050 patients: 1043 of these were wide local excisions (WLE), which included 11 level two oncoplastic procedures, with the remaining being excision biopsies. 957 of the WLEs were SNOLL procedures. ROLL was successful in 1065 (99.3 %) procedures. Of the eight ROLL failures, three were due to incorrect lesion localisation and two were due to diffuse ROLL signal. SLN biopsy was successful in 955 (99.8 %) of cases.

Conclusions

SNOLL/ROLL is a reliable, cheap and easy localisation technique where the marker cannot be dislodged during the procedure. Both localisations are performed on the day of surgery using the same equipment for signal detection in the breast and axilla.

Introduction

In recent years a range of techniques have become available for the localisation of impalpable breast lesions. Although historically wire localisation has been the predominant method used, non-wire techniques are becoming more popular for multiple reasons including patient comfort, logistical improvements and easier on-table localisation. Prior to the introduction of newer seed-based technologies using magnetic, radiofrequency and radar guidance, alternative non-wire techniques that were available included radioguided occult lesion localisation (ROLL), radioiodine seed localisation and intraoperative ultrasound, however adoption rates of these techniques were low: in the UK, of 98 breast units that responded to the iBRA-NET national practice questionnaire between October 2018 and April 2019, only 16 were not using wires as their main form of localisation [1]. Of these, nine were using Magseed, five were using ROLL and two were using radioiodine seed localisation.

Despite the benefits of the aforementioned seed-based technologies, the extra cost and requirement for specialist intraoperative equipment can prohibit their use. In addition, they are liable to dislodge during surgery. The ROLL procedure involves injection of Technetium 99 radioisotope into a breast lesion, allowing it to be detected with the same gamma probe that is used for radioisotope guided sentinel lymph node biopsy, whilst at the same time remaining in situ throughout the operation. The procedure was developed at the European Institute of Oncology in Milan in 1996 [2,3] and studies, including a meta-analyses, have shown it to be at least equivalent to wire-guided localisation [[4], [5], [6]]. A more simplified version of the procedure was introduced in our high-volume breast cancer unit (over 700 screen-detected and symptomatic breast cancer cases per year) as the predominant method for impalpable lesion localisation in 2006. It has been routinely used for both breast-conserving surgery and excision biopsies, including alongside radio-isotope guided sentinel lymph node biopsies (SNOLL).

The introduction of newer localisation methods has brought with it improved and more rigorous methods of surgical device evaluation [1]. We wanted to evaluate our outcomes using ROLL and SNOLL in light of this emerging data [7]. Here we present our 10 year experience using ROLL in a multitude of settings, demonstrating that it remains a versatile and effective tool for breast lesion localisation.

Methods

Study design

This was a single-centre retrospective cohort study. All patients aged 16 and over who underwent breast-conserving surgery or excision biopsy of a breast lesion using ROLL between 1/1/2008 and 1/1/2018 at Norfolk and Norwich University Hospitals were included in the study. A list of all patients undergoing breast surgery during this time period was obtained and data was extracted from electronic records for all those undergoing ROLL procedures. Cases were excluded from analysis if ROLL was planned but did not proceed. The study was registered locally as a service evaluation.

Procedures

On the morning of surgery or the afternoon before, 10 MBq ^99mTechnetium tin colloid was injected with an air tracer into a pre-agreed location in the breast under ultrasound guidance and a skin mark was applied. Localisation was usually into the centre of the impalpable breast lesion or at the clip site for lesions that were not visible on ultrasound at the time of injection either primarily or following neoadjuvant systemic therapy. In multifocal cases, radioisotope was either injected into one or two of the lesions, or into breast tissue lying between two lesions. In some multifocal cases two forms of localisation were used (usually wire and ROLL). If sentinel lymph node biopsy was planned, a further 10 MBq ^99mTechnetium tin colloid was injected into the periareolar skin farthest from the lesion. No check mammography is required. Intraoperatively, a Gamma probe was used to localise the injected site and sentinel node. For wide local excisions, specimens were orientated at three points using both silk marking sutures and clips at each point and the gamma probe was used to check the location of radioactivity in the specimen. 2-view specimen X-ray was performed of the breast excision specimens and a verbal report was phoned through to the operating surgeon by a consultant radiologist or radiographer. The decision for intraoperative cavity shaves was based on the specimen x-ray findings and the surgeon’s discretion and therefore not performed routinely for all patients. Sentinel lymph node biopsy was performed with single-tracer radio-isotope. Oestrogen receptor status was routinely assessed on all core biopsies of invasive disease. HER2 status was routinely assessed on core biopsies of invasive disease in patients under 70 years old and on clinical request in patients aged 70 years or over. All cases were discussed at a breast multidisciplinary team meeting both following core biopsy and surgery and recommendation for further surgery was guided by the unit’s prevailing margin guidelines at the time.

Outcomes

The primary outcome was whether the excision procedure (either wide local excision or excision biopsy) was successful. This was defined as the presence of the lesion or the clip (in cases of pathological complete response) in the surgical specimen. Secondary outcomes were successful ROLL (lesion excision was possible using ROLL guidance); satisfactory specimen x-ray report; requirement for intraoperative cavity shaves; weight of wide local excision specimen excluding cavity shaves (in grams); satisfactory margins on the main excision specimen; recommendation for re-operation; successful SLNB as defined by the presence of at least one lymph node in the specimen; total number of lymph nodes removed as part of SLNB; and positive SLNB (defined as presence of macrometastases in the SLNB specimen).

Statistical analysis

Continuous variables are reported as mean (standard deviation) if normally distributed and median (interquartile range) if not. Categorical variables are reported as frequency (percent).

Results

1083 ROLL-guided excisions were planned in 1060 patients, of which ten were excluded as they did not proceed with ROLL localisation (Fig. 1). In eight cases, the lesion or clip was not able to be identified on ultrasound imaging therefore a wire was inserted under stereotactic control, one patient did not require any localisation as the lesion was noted to be within a palpable haematoma and in one patient the radioisotope tracked up the needle during injection therefore wire localisation was used instead. In total 1073 ROLL-guided excisions in 1050 patients were included. The median age at the time of first surgery was 63 (range 25 to 89).

Fig. 1.

Flow chart of patients included in the study.

12 patients had bilateral ROLL-guided excisions, four patients had a second ROLL-guided excision at a future date within the study period and seven patients had two separate ROLL-guided excisions from the same breast at the same time. In cases where one patient had two ROLL localisations, these are treated as two separate cases if they were removed separately and one case if they were removed en bloc.

ROLL for wide local excision

1043 lesions that were B4 and above were targeted by ROLL for wide local excision. Clinico-pathological features are shown in Table 1 and Supplementary Table 1. One of the invasive lesions was not biopsied as it was radiologically suspicious and was in a breast with a biopsy-proven palpable cancer therefore the decision was made to perform a wide local excision of this lesion at the time of surgery. The majority of lesions were removed with a simple wide local excision, however ten (1.0 %) underwent therapeutic mammaplasty and one patient had a Grisotti flap (Table 1).

Table 1.

Clinicopathological features and type of surgery for patients undergoing ROLL-guided WLE.

	All ROLL WLEs n = 1043
Method of lesion detection Screening Symptomatic Incidental	795 (76.2 %) 218 (20.9 %) 30 (2.9 %)
Multifocal on preoperative imaging No Yes	987 (94.6 %) 56 (5.3 %)
Lesion size (mm)+ Mammogram Ultrasound Not reported/not done	12 (8 – 18) 10 (7 – 15) 316
Histological diagnosis on core biopsy Invasive ductal carcinoma Invasive lobular carcinoma Invasive tubular carcinoma Mixed carcinoma Papillary carcinoma Other invasive carcinoma DCIS/pleiomorphic LCIS B4 B3 Intraductal papilloma Not biopsied	712 (68.3 %) 85 (8.2 %) 24 (2.3 %) 18 (1.7 %) 15 (1.4 %) 103 (9.9 %) 82 (7.9 %) 3 (0.3 %) 1 (0.1 %) 1 (0.1 %)
Type of breast surgery Wide local excision Therapeutic mammoplasty Wide local excision with Grisotti flap	1032 (98.9 %) 10 (1.0 %) 1 (0.1 %)
Type of axillary surgery SLNB With ALND With ROLL-guided lymph node excision ALND ROLL-guided lymph node excision with ALND None	957 (91.8 %) 24 (2.3 %) 2 (0.2 %) 42 (4 %) 1 (0.1 %) 44 (4.2 %)

⁺values are median (i.q.r.). ALND – axillary lymph node dissection, DCIS – ductal carcinoma in situ, ROLL – Radioguided occult lesion localisation, SLNB – sentinel lymph node biopsy.

The wide local excision procedure was successful in 1038 cases (99.5 %). ROLL was unsuccessful in eight cases (0.8 %). Of these, five required repeat wide local excision: three due to localisation of an incorrect target; and two because the target lesion was not identified in the surgical specimen. In the remaining three cases the tumour was able to be localised intraoperatively using the skin mark or palpation. Further details of the unsuccessful ROLL cases are given in Table 2.

Table 2.

Details of failed ROLL cases.

Method of lesion detection	MMG size (mm)	US size (mm)	Specimen X-ray	Intraoperative cavity shaving taken	Lesion in main specimen	Tumour in shave	Specimen weight (g)	Target lesion histology	Reason for ROLL failure	Second operation
Screening	10	9	Clip/lesion not seen	Yes	No	Yes (all of tumour)	Unknown	IDC	No tumour on specimen X-ray. Subsequent lateral shave taken and contained palpable tumour	No
Screening	12	7	Satisfactory	Yes	No	No tumour	37.0	IDC	Incorrect lesion targeted (intramammary lymph node)	Repeat ROLL WLE
Screening	Unknown	6	Satisfactory	No	No	N/A	47.0	IDC	No clear reason – spiculated mass seen on specimen x-ray but no tumour or biopsy site on histology, just small amount of DCIS	Repeat WLE (no ROLL)
Screening	6	6	Clip/lesion not seen	Yes	No	No tumour	6.0	IDC	Diffuse signal	Repeat ROLL WLE
Symptomatic	0	8	Satisfactory	No	No	N/A	8.5	IDC	Incorrect lesion targeted (intramammary lymph node)	Repeat ROLL WLE
Screening	13	0	Not satisfactory	Yes	Yes (2 foci of DCIS up to 4 mm and biopsy site)	No	21.5	DCIS	Very diffuse signal over upper breast ?due to haematoma. Excision based on skin marking. Thick cavity shaves taken.	No
Symptomatic	12	17	Clip/lesion not seen	Yes	No	Yes (all of tumour)	65.0	IDC	WLE performed according to skin mark and maximum radioactivity - no specimen on X-ray or US. Further WLE (14 g) from deep central portion as cancer became palpable but not radioactive	No
Screening	N/A	9	Indeterminate	No	No	N/A	9.0	IDC	Incorrectly targeted lesion and specimen X-ray difficult to interpret	Repeat WLE (wire-guided)

DCIS – Ductal carcinoma in-situ, IDC – invasive ductal carcinoma, MMG – mammogram, N/A – not applicable, US – ultrasound.

The median specimen weight of the main specimen was 43 g (i.q.r. 29–61 g). 812 (78.0 %) of patients had a satisfactory intraoperative specimen x-ray and 260 (25.0 %) underwent intraoperative cavity shaves (Table 3). Of those who had intra-operative shaves, 40 (15.4 %) had invasive or in situ disease in the shave. 149 (14.3 %) patients underwent a second breast operation and 25 (2.4 %) of these underwent a third breast operation (Table 3). A further five patients were recommended either a margin re-excision or mastectomy after the initial operation due to unsatisfactory margins, however they declined further surgery.

Table 3.

Surgical outcomes following ROLL WLE.

	All ROLL WLEs n = 1043
Specimen x-ray report Satisfactory Not satisfactory Clip/lesion not visible Margins unsatisfactory Indeterminate Not done / Report unknown	812 (78.0 %) 199 (19.1 %) 17 (1.6 %) 182 (17.5 %) 8 (0.8 %) 24 (2.3 %)
Specimen weight Specimen weight (g)+	43 (29–61)
Histology Invasive carcinoma DCIS Pathological complete response No residual disease (lesion excised with core biopsy) Tumour not found	942 (90.3 %) 62 (6 %) 30 (2.9 %) 4 (0.4 %) 5 (0.5 %)
Lesion size Total lesion size including in situ component (mm)+	14 (10–21)
Radial margins on main excision specimen Close (<1 mm) Negative (≥1 mm)	186 (17.8 %) 857 (82.2 %)
Intra-operative cavity shave Yes No	260 (25.0 %) 783 (75.0 %)
Invasive or in situ disease in shave Yes No	n = 260 ,220 (84.6 %) 40 (15.4 %)
Re-excision required Yes No	154 (14.8 %) 889 (85.2 %)
Second operation on breast Repeat wide local excision Margin re-excision Mastectomy	n = 150 5 (0.5 %) 122 (11.7 %) 23 (2.2 %)
Third operation on breast Margin re-excision Mastectomy	n = 25 3 (0.3 %) 22 (2.1 %)
SLNB successful Yes No	n = 957 ,955 (99.8 %) 2 (0.2 %)
Number of SLNB nodes Number of nodes removed as part of SLNB+	n = 957 2 (1–3)
Positive SLNB Yes No	n = 957 ,139 (14.5 %) 818 (85.5 %)

⁺values are median (i.q.r.). For multifocal disease information is given about the largest tumour.

SNOLL

957 patients underwent SNOLL, of whom only two had no lymph nodes retrieved (Table 3). The median number of lymph nodes removed was 2 (i.q.r. 1–3) and 139 patients (14.5 %) had macrometastases in the sentinel node(s). Three patients underwent ROLL-guided excision of a lymph node simultaneously with SNOLL. 25 patients proceeded to axillary lymph node dissection in the same operation as SLNB due to either clinical concern, positive SLN status assessed using intraoperative touch imprint cytology (2011 to 2012) or participation in a local audit (2016 to 2017).

ROLL-guided breast lesion excision

30 patients underwent ROLL-guided excisions (Supplementary Table 2). Six (20 %) of these were in patients who had a tumour in the ipsilateral breast, of which four (13.3 %) patients had two separate ROLL injections for the excision biopsy and the wide local excision. ROLL was successful in all cases. Three (10 %) patients had intraoperative cavity shaves, of which one contained columnar cell change with atypia. Further excision was not required for any case.

Discussion

This is one of the largest series of ROLL/ SNOLL cases published to date, consisting of 1043 wide local excisions and 30 breast excisions, including complex cases such as multifocal disease and ROLL-guided lymph node excision with radioisotope-guided SLNB. Our data shows that ROLL is a very effective method for guided WLE with comparable lesion localisation (99.4 %) and re-operation (14.3 %) rates to magnetic seed (99.8 %; 12.3 %) and wire localisation (99.1 %; 13.2 %) [7], particularly taking into consideration that margin guidelines were less stringent during the more recent time period of the study by Dave et al. Intraoperative cavity shaves were performed less frequently (25 % vs 35.7 % and 35.5 % for magnetic seed and wire, respectively). The margin re-excision rate was lower than the UK average of just under 19 % between 2014 and 2017 according to GIRFT data [8]. We did measure weight of the removed specimen. Dave et al. did not report specimen weights, however our median specimen weight of 43 g was comparable to the mean weights reported elsewhere for Magseed- and wire-guided wide local excisions (39.6 g and 44.5 g, respectively) [9]. Due to the retrospective nature of our study, we did not include information about post-operative complications and readmissions as this information was not readily available electronically.

Concomitant SLNB (SNOLL) showed excellent identification of sentinel lymph nodes, possibly due to both peritumoral and peripheral injection of radioisotope. Furthermore, the median number of lymph nodes retrieved was the same as has been previously reported with isotope alone [10], suggesting that the dual injection did not increase the number of “hot” lymph nodes. In addition, isotope injection for SLNB did not interfere with identification of the index lesion using ROLL, or even, where it was required, ROLL-marked lymph nodes, suggesting that a radio-isotope only technique is also suitable for targeted axillary dissection. The SNOLL technique has previously been reported, however those studies use different, more complex methodology such as using different sizes of human serum albumin macroaggregate bound to ^99mTc for the breast and sentinel node; and routinely performing lymphoscintigraphy of the breast and axilla, which required a delay of 3 h after the injection [11]. In two cases in our study, the signal was too diffuse, possibly due to underlying haematoma, and this may have been identified preoperatively if scintigraphy had been performed.

Ongoing studies are developing and evaluating different localisation techniques, many of which are favouring seed-based technologies due to the ability to temporally uncouple localisation from surgery, as well as reducing the need for two radiological interventions in neoadjuvant cases by placing the seed instead of a clip. However for patients not undergoing chemotherapy, routine seed placement at the time of biopsy needs to be done judiciously in order to avoid unnecessary seed insertion and its associated patient and financial costs. A totally magnetic technique for lesion localisation and sentinel lymph node biopsy has been reported, which uses a single probe and has the benefit of additional flexibility with the timing of the sentinel node injection, although this can cause MRI artefacts [12].

The benefits of ROLL are that the same probe and injection material can be used for the breast and axilla; the probe is slim, can be used with other surgical instruments and has directionality; the marker location cannot be dislodged during the procedure; and it can be used for multifocal procedures, although it is difficult to differentiate between two injection sites if they are <1 cm apart. There are also potential environmental benefits of using a single, reusable probe and no seeds as this reduces the amount of single-use material required. The drawbacks are the radiation safety and costs associated with the use of radioisotope (although this is less than radioactive seeds); lack of localisation confirmation at the time of injection unless scintigraphy is used; and the half-life of radio-isotope, which limits timing of the injection to the day before or morning of the procedure. As the SLNB injection is done at the same time as the ROLL injection this has not caused undue scheduling difficulties for patients undergoing concomitant SLNB. Furthermore, due to the large catchment area of our unit, many patients prefer to avoid an additional hospital visit for seed localisation prior to the day of the operation. We have only used ultrasound-guided ROLL insertion, therefore a clip needs to be in situ if the lesion is ultrasound-occult, however stereotactic ROLL injection is technically possible [11].

In our unit we maintain two localisation techniques: ROLL and a seed-based technique (radio-frequency identification tags [13] or implantable reflectors), in order to have access to the strengths of each. This has allowed the workflow to continue even at times when there has been a shortage of either radioisotope or seeds. In general, ROLL is preferred if there is a short time from booking to surgery (and therefore insufficient time for the seed to fix into place) and for lesions where there is concern about the seed dislodging easily during surgery e.g. those on the pectoralis major muscle, very superficial lesions, solid lesions or small lesions; whilst seeds are preferred for patients expected to undergo breast conservation surgery after neoadjuvant therapy who do not require MRI, or lesions that have not already been clipped.

Chu et al. reported that the radiation rose associated with ROLL is very low. Effective dose for a patient, which is 9.25 µSv, less than half the dose of a chest X-ray (0.02 mSv). Radiation due to additional mammograms needed for wire guided or other seed localisation (1–2 mSv) may exceed that involved for 100 to 200 ROLL procedures. Finger doses to breast surgeons and radiologists are also minimal, amount to 9.3 mSv and 0.5 mSv, respectively. For a surgeon performing 100 procedures per annum, a finger dose of 1 mSv is received, which is well below the annual limit of 150 mSv. Effectively no contamination is detected in wastes and porters receive no radiation. Therefore no additional protection measures are required. Roll technique results in less exposure to radiation compared to other techniques requiring a check mammogram [14].

ROLL is a reliable, safe, cheap and easy technique with a very high success rate that can be used to localise breast lesions alone or with SLNB (SNOLL). However, the retrospective design and lack of direct comparison with alternative localisation techniques limit broader generalisability and assessment of relative oncological or patient-reported outcomes. Future prospective designed work can address these weakness and add further clarification of various localisation techniques.

CRediT authorship contribution statement

Bahar Mirshekar-Syahkal: Writing – original draft. Haifa Alotaibi: Data curation. Sendhil Rajan: Data curation. Mathew Gray: Methodology. David Newman: Methodology. Maged Hussien: Writing – review & editing, Supervision, Project administration, Investigation, Formal analysis, Data curation, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.