Article Text

Download PDFPDF

Sentinel lymph node biopsy in breast cancer patients after overnight migration of radiolabelled sulphur colloid
  1. N Lamichhane1,
  2. K W Shen2,
  3. C L Li3,
  4. Q X Han2,
  5. Y J Zhang4,
  6. Z M Shao2,
  7. Z Z Shen2
  1. 1Department of Surgical Oncology, B P Koirala Memorial Memorial Cancer Hospital, Bharatpur, Chitwan, Nepal
  2. 2Department of Surgical Oncology and Breast Surgery, Fudan University Cancer Hospital, Shanghai, Peoples’ Republic of China
  3. 3Department of Ultrasonology, Fudan University Zhong Shan Hospital, Shanghai, Peoples’ Republic of China
  4. 4Department of Nuclear Medicine, Fudan University Cancer Hospital, Shanghai, Peoples’ Republic of China
  1. Correspondence to:
 Dr Nirmal Lamichhane
 Department of Surgical Oncology, B P Koirala Memorial Memorial Cancer Hospital, Bharatpur P O Box 34, Bharatpur, Chitwan, Nepal;


Objective: To evaluate the performance and feasibility of sentinel lymph node biopsy in breast cancer patients using technetium-99m (99mTc) sulphur colloid and gamma probe.

Methods: From May 2000 to March 2001, 70 patients with a tumour less than 5 cm with clinically negative axillary lymph nodes underwent sentinel node biopsy followed by standard axillary dissection. 99mTc sulphur colloid was injected around the primary tumour the day before surgery and a gamma probe was used to detect the sentinel lymph node during the surgical procedure. Sentinel lymph node biopsy was compared with standard axillary dissection for its ability to accurately reflect the final pathological status of the axillary nodes.

Results: The sentinel lymph node was successfully identified in 67 of 70 patients (95.71%). The number of sentinel lymph nodes ranged from 1–5 (mean 1.5) and non-sentinel nodes ranged from 5–22 (mean 13.3). Of the 67 patients with successfully identified sentinel lymph nodes, 43.28% (29/67) were histologically positive. Sensitivity of the sentinel lymph node to predict axilla was 82.75%; specificity was 100%. Positive and negative predictive values were 100% and 88.3% respectively. The sentinel lymph node was falsely negative in five patients, yielding an accuracy of 92.53%. Sentinel lymph node biopsy was more accurate for T1 tumours than for T2 tumours.

Conclusions: The gamma probe guided method after overnight migration of 99mTc sulphur colloid is technically feasible for detecting sentinel lymph nodes in most breast cancer patients, accurately predicting the axillary lymph node status, and appears more accurate for T1 lesions than for larger lesions. This minimally invasive axillary staging procedure represents a major advance in the surgical treatment of breast cancer.

  • H&E, haematoxylin and eosin
  • KP-1, Klebsiella pneumoniae
  • 99mTc, technetium-99m
  • breast cancer
  • sentinel lymph node
  • biopsy

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

The purpose of axillary node dissection in breast cancer patients is to achieve information of disease stage, obtain local control, and assist in the decision making for adjuvant therapy. However, for those patients with pathologically negative nodes, dissection does not increase survival.1 About 60%–70% of patients with early breast cancer have no metastasis to the regional lymphatic basin,2,3 so axillary dissection and consequent morbidity can be avoided if there is an accurate alternative for axillary node staging. Sentinel lymph node biopsy may provide physicians with the ability to distinguish negative from positive axillary lymph nodes in a relatively simple fashion.

In 1992, Morton et al reported this technique in patients with clinically node negative melanoma.4 This technique has now been defined for breast carcinoma by Krag et al5 and Giuliano et al6 as a rational and accurate approach to the study of the regional lymphatic basin. Sentinel lymph nodes are physiologically determined as entrance lymph nodes to the lymph node basin and drain a particular tumour and, if negative, predict the status of the remaining axillary nodes.7,8 During the past eight years, sentinel lymph node biopsy in breast cancer patients has become an exciting research topic. Results from international breast cancer centres show that, with the use of optimal techniques, the accuracy of sentinel lymph node biopsy to predict axillary nodal status is 95% or more. Using radiolabelled sulphur colloid as a tracer, after its overnight migration, we performed sentinel lymph node biopsy in patients with breast cancer and evaluated its feasibility for Chinese patients.



Seventy breast cancer patients with clinical TNM status T1–2, N0 admitted to our hospital were included in a prospective study performed between May 2000 and March 2001. Both mastectomy and breast conservation patients were equally eligible. Informed consent for injection of radiotracer around their breast mass and surgical procedure to be performed was taken from each patient undergoing the study. The institutional ethical review board at Fudan University Medical Center had accepted the procedure. The age of the patients ranged from 30 to 86 years with a mean age of 50.9 years; 38.67% were premenopausal and 61.4% were postmenopausal. All patients underwent sentinel lymph node biopsy and consecutive axillary dissection with the preplanned mode of surgery.


Ten to 16 hours before the operation, 1–2 mCi (1 mCi = 37 MBq) of unfiltered technetium-99m (99mTc) sulphur colloid (prepared by the department of nuclear medicine, Fudan University Cancer Hospital) in a volume of 3–5 ml was injected in divided aliquots into the breast tissue surrounding the primary tumour or biopsy site at the 3, 6, 9, and 12 o’clock positions. Injections were made at 1–3 mm surrounding the tumour or biopsy cavity. Breasts soft in consistency and bigger in size received the larger volume of injection. Among 70 cases, only a single patient presented with itching and erythema of skin at the injection site, which subsided after antihistaminic administration. A hand held gamma probe (Capintec Gammed IV, Capintec Inc, New Jersey, USA) was used to detect the distribution of radioactivity around the breast and axilla before the operation. Hot spots were detected and the skin marked accordingly. A hot spot was defined as the spot that had the greatest radioactivity counts in the lymphatic basin, which was at least 25 counts per 10 seconds or greater. After resection, the activity of the resection bed should be less than 10% of the most radioactive node.7,8 In 63 patients, static lymphoscintigraphy was performed before operation to ascertain the overall distribution of the radiotracer. Skin marks were made over the hot spot to facilitate the search during the operation.

For sentinel node biopsy, a 4–5 cm skin incision was made over or near the hot spot, and a hand held gamma probe was used to guide the further dissection until the hot node resection was complete. In 27 cases, sentinel node biopsy was performed after the dissection of skin flaps. In the four most preliminary cases, the sentinel node was dissected from the specimen after radical mastectomy. A standard axillary dissection was then performed. When the sentinel lymph node could not be tracked even in the dissected axillary specimen, the procedure was considered to be a failure.

Pathological examination

Excisional biopsy of the lesion was performed and frozen sections taken for the confirmation of cancer. The sentinel node biopsy specimens were identified as SLN1, SLN2, etc and were separately infiltrated into a 10% solution of formalin before delivery. All lymph nodes in the specimen were identified and dissected from the surrounding tissue by the pathologist. The nodes were then bisected and embedded for sectioning. One or two sections obtained from the central cross section from each block were then stained with routine haematoxylin and eosin (H&E) for the detection of cancer cells.

In 63 sentinel lymph nodes of 43 cases which showed no cancer cells by conventional pathological method, further evaluation was performed by using immunohistochemical stains for cytokeratin (antibodies for CK8 and CK19, DAKO products).9 Three sections were taken of the paraffin block, two stained for cytokeratin and one with H&E. Staining was performed in the department of pathology, Fudan University Cancer Hospital. A positive finding was defined as a cluster of cells in brown staining in its cell membrane or plasma which had to be further verified in H&E staining of the same level. Scattered anaplastic cells in the subcapsular region of lymph node which was positive for cytokeratin antibody were evaluated with immunohistochemical stains for Klebsiella pneumoniae (KP-1) (CD-68) to exclude histiocytes.10

Statistical analysis

False negative rate was defined as the percentage of patients who had histologically negative sentinel lymph nodes but other nodes positive. Accuracy was defined as the percentage of patients in which the sentinel node status has accurately represented the lymph node status of patient. Fisher’s exact test and χ2 tests were used to analyse the impact of different factors to this method.


The clinical data of patients and pathological characteristics of the tumour are detailed in tables 1 and 2. Among the 63 patients who underwent lymphoscintigraphy, all had lymphatic drainage towards the axilla, including 12% who had simultaneous internal mammary drainage.

Table 1

 Clinical data of patients (n = 70)

Table 2

 Pathological features of the tumour (n = 70)

The sentinel lymph node was successfully detected in 67 of the 70 patients—that is, a detection rate of 95.72%. In three cases that were failures in detecting the sentinel lymph node, including one case in which the pathologist could not identify a lymph node-like structure in the surgically sent specimen. These three patients were 58, 62, and 60 years old with a tumour size of 2, 3, and 2.3 cm respectively.

In 67 successful cases, the number of nodes detected ranged from 1–5. A total of 101 sentinel nodes were retrieved and the mean number of sentinel lymph nodes per case was 1.5. The number of dissected non-sentinel nodes ranged from 5–22, a total of 925 nodes, with a mean of 13.8 nodes per case. Except for one case in whom the sentinel lymph node was in the third intercostal space, all the other sentinel nodes were located at level I of the axilla.

A total of 29 (43.3%) patients had cancer cells in their lymph nodes, with sentinel nodes positive in 24 cases (sensitivity 82.75%). In the remaining 38 patients, the sentinel and non-sentinel nodes were all negative (specificity 100%). Sentinel node biopsy missed metastatic lymph nodes in five patients (false negative rate 7.46%). The positive predictive value was 100%, negative predictive value was 88.37%, with an overall accuracy of 92.53% (see table 3).

Table 3

 Summary of sentinel node biopsy

Immunohistochemistry for the cytokeratins CK8 and CK19 in 65 H&E staining negative sentinel lymph nodes showed a few single anaplastic cells in five lymph nodes that were considered to be a histiocytes (non-cancer cells) by the pathologist; this was reconfirmed because these cells also stained positive for KP-1.

When the tumour size was stratified, the accuracy was 100% with no false negatives in 25 cases with a tumour size less than 2 cm. In nine cases the patients had undergone excision or incision biopsy; there was also no failure or false negative result. In seven cases the sentinel node was the only positive node, and of these there were four patients who had ductal carcinoma in situ with early invasion (table 4). Our success rate reached 100% after 20 biopsy procedures (fig 1).

Table 4

 Stratification of patients (n = 70) and tumour characteristics and respective success rates, accuracy, and false negative rate (FNR)

Figure 1

 Learning curves for the 70 patients enrolled in the study. They were divided into seven groups with 10 patients in each group. Success rate and false negative rate were calculated for each increment of 10 patients in a cumulative fashion that yielded this curve.


The sentinel lymph node is the first cancer cell draining node.4,5 Theoretically, there is a gradual spread of tumour cells from the sentinel nodes to other nodes so further spread of cancer cells may be predicted by the sentinel nodes.4 In cutaneous melanoma, the accuracy of sentinel node biopsy to predict lymphatic basin involvement is over 95%.11 In western countries, this is already a part of regular treatment protocol. In the same way, the sentinel lymph node biopsy procedure in breast cancer patients is also a regular procedure in many western countries. Current literature relating to sentinel node biopsy suggests that the accuracy of sentinel nodes to predict axillary status is greater than 95%, with a false negative rate ranging from 0 to 5%.

In this study, our success rate using radiotracer and gamma probe is 95.7%. Results of this preliminary study are similar to the other preliminary studies of Krag et al (82%),5 Pijpers et al (92%),12 and Borgstein et al (94%)13 using similar methods. Giuliano et al’s landmark report of a 65% success rate with blue dye shows a success rate of 58.6% in the first half of study and 72.4% in the second half, with a significant increase in success rate in the second half of the study.6 More recently, these authors reported successful mapping in 93% and 99% of cases.14,15 We experienced a 91.4% success rate in first half of study and all the failures occurred in first 20 cases. In the second half, the success rate was 100%. This reflects the developmental stage of a new technique, and is shown in the discrete learning curve (fig 1).16 We also experienced a faster and accurate detection and dissection of sentinel node specimen in the later cases of our study. However, our detection rate was lower compared with the most recent reports of experienced doctors. We believe that we are now in the learning stage and think this is the main reason for the difference. Some reports suggest that blue dye and isotopes are complementary, and that sentinel node identification is facilitated when the techniques are used together, thereby accelerating the learning curve.17,18 As the blue dye is not commercially available in the Peoples’ Republic of China, we have no experience using dye or dual method.

In our study, the observed accuracy of sentinel nodes for the detection of metastatic disease in permanent sections was 92.5%, and false negative cases were 7.5%. The false negative rate is of great concern when this procedure is applied as a treatment protocol. O’hea et al, in a series of 60 sentinel node biopsy procedures using dye plus isotope localisation, found that all of their three false negative sentinel nodes occurred in tumours larger than 1.5 cm.18 Veronesi et al have similarly found no single false negative sentinel lymph nodes in tumours less than 1.5 cm in diameter.19 In this study, we also found that the false negative rate was increased and accuracy diminished with increase in tumour size (table 4). This inaccuracy in larger tumours may be caused by alternate lymphatic drainage pathways or simply by the prevalence of axillary metastasis in patients with larger tumours. In our series, the false negative rate is high compared with that of more experienced doctors. In our trial, T1 lesions only consisted of 37% of all the cases and the high false negative rate might be a function of the greater percentage of larger lesions.

For the methodology, the tracer injection timing is an issue to be debated. The physical half life of 99mTc sulphur colloid is six hours. Krag et al reported the average time between injection and surgery as being 2.9±1.9 hours.7 This means that surgery is carried out in the afternoon only, which is inconvenient for both patients and surgeons. When the tracer is injected a day before the surgery, the radioactivity in the node, the axillary bed, and the background will certainly be markedly decreased on the day of surgery. But our experience showed the remaining radioactivity in the sentinel lymph nodes, however, was strong enough to be detected by the gamma probe. All the nodes exceeded the criteria of greater than 25 counts per 10 seconds on the day after injection. Static lymphoscintigraphic studies also showed a hot spot at the lymphatic basin after two or more hours and it remained the most active node for the next 24 hours. So the procedure is possible without any problems using overnight migration of 99mTc sulphur colloid radiotracer. This method gives us enough time to perform lymphoscintigraphic studies of the breast carcinoma, and the low amount of radioactivity is more acceptable to surgeons, anaesthesiologists, nurses, and pathologists.

The multiple injection of radiocolloid around the tumour may give the best diffusion of tracer around the tumour, which in turn is carried to the most accurate sentinel node. Although some authors have used subdermal, subcutaneous,19 or periareolar20 injection of radiotracer yielding an excellent sentinel lymph node identification rate, there remained a question as to whether the identified flow was a true representation of the lymphatic flow of the tumour bed.21 The radiocolloid was injected around the breast tissue surrounding the primary tumour or biopsy site, taking the size and consistency of the breast into consideration. In smaller breasts with a hard texture, the volume injected was smaller, and this is the most frequent situation for our patients.

To determine the factors influencing the success rate and accuracy, McMasters et al, in a multicentre study including 806 patients, analysed the age, site of tumour, size of tumour, type of biopsy method, and experience of the surgeon. They found that the success rate was low in patients older than 50 years (p  =  0.03).22 The false negative rate was high (p = 0.004) for outer quadrant tumours, which is similar to the report of Krag et al (p = 0.05).7 In our series, the two real failures were in patients older than 50 years (58 and 62 years) and, in five false negative cases, the lesions were larger than 2 cm at their greatest diameter and all were located in the outer quadrant of the breast. Since the sample size was too small, our data did not show a statistically significant result (table 4). The probable reason for older patients to have a higher failure rate might be a decrease in the capacity of lymph nodes to retain the radioactive colloid, because lymph nodes are replaced by fat in elderly persons.7 In upper outer quadrant tumours, difficulty in discriminating the signal from background after peritumoural injection of the radiotracer may be the cause of a low detection rate; this is also considered to be a drawback of the radioguided procedure.15

So far as radiation hazard to operating room personnel, pathologists, and equipment during a breast sentinel node biopsy is concerned, Stratmann et al reported that with 0.7–1.0 mCi of 99mTc sulphur colloid injected 1.5–3 hours before the sentinel node biopsy, primary surgeon, scrub nurse, and pathologist can perform respectively 2190, 33 333, and 14 705 hours of procedural work before surpassing Occupational Safety and Health Administration limits. Operating instruments, pathology slides, and cryostats do not require special handling.23 Only the lumpectomy specimens should be quarantined until the dose rate equals background levels. Although we injected 1–2 mCi of 99mTc sulphur colloid, we performed surgery 10–16 hours after injection. Twice the half life of 99mTc had elapsed before surgical procedures as its half life is only six hours. Compared with the report of Stratmann et al, the exposure to operating room personnel and the pathologist performing frozen section is quite low. Exposure to a patient from a single procedure is of no meaningful concern. In general this procedure requires one 60th the dose of technetium delivered during a routine bone scan. Technetium is a safe radiopharmaceutical because of its relatively low energy gamma ray (140 kEV), short half life, and lack of beta emission.24

Our hospital breast cancer database shows that 23% of all the breast cancer patients in our department had T1 tumours and 56% had T2 tumours. Only 19% of cases of T1 lesions were clinically examined to have axillary lymphadenopathy, and 29% of cases were found to have a pathologically positive axilla. Currently, these patient groups are being treated with the traditional method of axillary dissection. The sentinel node biopsy will establish itself as a safe and effective alternative to axillary dissection in this patient subgroup.25 This modification will decrease the operative morbidity of surgical therapy for breast cancer, providing accurate staging and more individualiesd therapy. This minimally invasive axillary staging procedure is a major advance in the surgical treatment of breast cancer. As radiocolloid can be injected a day before, there will be enough time for lymphoscintigraphy and surgical preparation.


We are very grateful to Dr Chen Zhong Wei, MD, pathologist for sincere help in preparation and reading our permanent slides of the sentinel lymph nodes. Many thanks to Miss Du Hui Feng, BSc, nuclear medicine technologist for her preparation of technetium-99m labelled radiocolloids for injection.