The Role of Axillary SUV_max in ¹⁸F-FDG PET/CT in Predicting the Number of Axillary Metastases of Breast Cancer

By Ufuk Karabacak¹, Halil Turkan², Gokhan Coskun³, Murat Can Mollaoglu¹, Zekiye Hasbek⁴, Kursat Karadayi⁵

Affiliations

1. Department of Surgical Oncology, Sivas Numune Hospital, Sivas, Turkiye
2. Department of Surgical Oncology, Sanliurfa Mehmet Akif Inan Training and Research Hospital, Sanliurfa, Turkiye
3. Department of Surgical Oncology, Tokat State Hospital, Tokat, Turkiye
4. Department of Nuclear Medicine, Faculty of Medicine, Cumhuriyet University, Sivas, Turkiye
5. Department of Surgical Oncology, Faculty of Medicine, Cumhuriyet University, Sivas, Turkiye

doi: 10.29271/jcpsp.2023.04.374

ABSTRACT
Objective: To investigate the role of positron emission tomography/computed tomography (PET-CT) in determining the maximum number of axillary lymph node metastasis (ALNM) detectable in sentinel lymph node biopsy (SLNB).
Study Design: Observational study.
Place and Duration of the Study: Sivas Cumhuriyet University Faculty of Medicine, Turkiye, from January 2015 to August 2021.
Methodology: A total of 104 breast cancer patients who underwent surgery after a PET-CT scan were examined. A receiver operating characteristic (ROC) analysis was utilised to determine optimal cut-off values for the standardised uptake values of the primary tumour (pSUV_max) and axillary lymph nodes (nSUV_max) in the presence of ALNM and the presence of more than two ALNMs.
Results: The presence of more than two ALNMs was associated with pSUV_max, nSUV_max, LVI, and the number of LNs detected on PET-CT. In the ROC analysis, for the ability to predict more than two ALNMs in SLNB/axillary lymph node dissection (ALND), cut-offvalues were calculated as 4.65 for pSUV_max (AUC=0.669, sensitivity=66.7%, specificity=62%, PPV=0.482, NPV=0.800, p=0.006) and 1.75 for nSUV_max (AUC=0.838, sensitivity=81.8%, specificity=88.7%, PPV= 0.676, NPV=0.913, p<0.001).
Conclusion: Low sensitivity, NPV, and accuracy values that limit the use of PET-CT in preoperative axillary evaluation can be increased by targeting the criterion of more than two ALNMs. Thus, PET-CT can be used more effectively in axilla management.

Key Words: Breast cancer, Positron emission tomography, SUV_max values, Axillary lymph node, ACOSOG Z0011.

INTRODUCTION

Breast cancer is the most common cancer and the leading cause of death in women.¹ Treatment of breast cancer is multidisciplinary and initial staging is critical for an appropriate treatment plan. One of the many methods used for staging is positron emission tomography (PET-CT). Besides local staging, PET-CT has the advantage of scanning the whole body and detecting extra-axillary lymph nodes (LN) bone and distant organ metastases.² PET-CT is based on the principle of cancer cells having a more active glucose metabolism than other tissues. To assess the glucose metabolism of tissues, the uptake of the [18F] fluorodeoxyglucose ([18F] FDG) molecule is measured, the standardised uptake value (SUV) is calculated, and the primary tumour and its metastases are evaluated.^3-5

Today the primary goal of breast cancer surgery is to conserve the breast and avoid axillary lymph node dissection (ALND) if possible. Sentinel  lymph  node  biopsy  (SLNB)  has  become  a  routine  procedure  in  the  historical  development  of  surgery  to  limit  the  indication  of  axillary  dissection  and  to  reduce  the  associated  morbidities.⁶ Recent studies have shown that ALND can be avoided by considering the number of metastatic LNs detected and some accompanying criteria, even if there is metastasis in SLNB.^7,8 Thus, the number of metastatic lymph nodes gained importance in axilla management. The  aim  of  this  study  was  to  investigate  the  role  of  the  axilla  SUV_max  value  in PET-CT  and  the  number  of  possible  metastatic  lymph  nodes.

METHODOLOGY

Patients who were operated for breast cancer in Cumhuriyet University Surgical Oncology Clinic, between January 2015 and August 2021 were analysed retrospectively. Demographic findings, operative notes, stages, pathology, and PET-CT results were searched from clinical records. Exclusion criteria were the absence of preoperative PET-CT or the inaccessibility of PET-CT images, having undergone diagnostic surgical excision before PET-CT, having a different cancer history, and receiving neoadjuvant chemotherapy. One hundred and four consecutive patients who met these criteria were included in the study.

All surgeries were started with 5cc isosulfan blue injection for SLNB. It was aimed to sample at least three LNs during SLNB. ALND was not applied in the absence of metastasis. The criteria considered for ALND were detection of metastases in frozen section examination of sentinel lymph node (SLN), failure to identify SLN, and detection of metastases in unstained but suspicious-looking LNs. Thus, the pathological evaluation of the axilla was optimal. For each patient, primary tumour (pSUV_max) and axillary LN (nSUV_max) SUV_max values in preoperative PET-CT; the number of LNs detected on PET-CT (those without FDG uptake or pathological appearance are also included.); the types of surgery applied to the breast and axilla; pathological type; the number of LN metastases detected during SLNB or ALND; histologic grade; T stage; N stage; lymphovascular invasion (LVI); perineural invasion (PNI); Ki-67; estrogen receptor (ER); progesterone receptor (PR); and HER-2 status were examined.

PET/CT imaging was performed with a combined PET/CT scanner (Discovery 600 PET/CT GE Medical Systems, USA). Each patient fasted for at least six hours before imaging. After ensuring that blood glucose was <180 mg/dL, approximately 0.14 mCi/kg ¹⁸F-FDG was administered intravenously one hour before image acquisition. Attenuation correction of PET images was performed with the CT data. The CT scan was performed first and right after the CT acquisition, a standard PET imaging protocol was taken from the cranium to the mid-thigh with an acquisition time of three min/bed in 3-D mode. All PET studies were acquired in 3-D mode. CT images were acquired with 70 mA, 120 kV, and an axial slice thickness of 2.5 mm. CT and PET images were matched and fused into transaxial, coronal, and sagittal images. The data were transferred via the Digital Imaging and Communications in Medicine protocol to a processing workstation (AW Volume Share5 GE Medical Systems S.C.S, France). Then, the visual and semi-quantitative analyses were performed, respectively. For PET images, an adaptive threshold setting of 42% of the maximum lesional metabolic activity was used, and the ROI was placed within the tumour while avoiding the peripheral area. The standardised uptake value (SUV) was calculated using the following formula: [Activity of ROI (mCi / ml) × Bodyweight (gram)] ÷ Injected dose (mCi).

The Kolmogorov–Smirnov test was used to compare the distribution of random sample. A chi-square test was used to compare categorical variables. The independent samples t-test was used to compare the normally distributed data, and the Mann-Whitney U-test was used to compare the data that were not normally distributed. ROC analysis was used to determine the cut-off values ​​of pSUV_max and nSUV_max for any number of axillary lymph node metastasis (ALNM) and more than two ALNMs. Binary logistic regression analysis with univariate and multivariate models was used to examine the risk factors affecting more than two ALNMs. Backward Wald method was used to include independent risk factors in the multivariate model. Analysis results were presented as mean±standard deviation and median (minimum–maximum) for quantitative data and frequency (percent) for categorical data. Data were analysed with IBM SPSS V23. All p-values ​​lower than 0.05 were considered statistically significant.

Figure 1: Axillary Lymph Node Metastasis Prediction Performance of PET-CT. (A) ROC curve of pSUV_max and nSUV_max to predict the presence of any number of ALNM. (B) ROC curve of pSUV_max and nSUV_max to predict the presence of more than 2 ALNMs. pSUV_max: Primary tumour maximum standardised uptake value, nSUV_max: Nodal maximum standardised uptake value, ALNM: Axillary lymph node metastasis, PET-CT: Positron emission tomography, ROC: Receiver operating characteristic.

RESULTS

The characteristics of the enrolled patients and the association with ALNM are summarised in Table I. ALNM was detected as a result of SLNB or ALND in 62 (59.6%) of 104 patients included in the study. Thirty-three patients (32%) had more than two ALNMs. The median value of pSUV_max without ALNM was 3.8 (1-16), while the median of those with ALNM was 5 (1.9-18, p=0.042). The median value of nSUV_max without ALNM was 1.0 (1-13.8), while the median of those with ALNM was 2.2 (1-14.7, p<0.001).

In the univariate model, it was shown that the risk of having more than two ALNMs during SLNB or ALND increased with the increase in nSUV_max and pSUV_max values, being grade 2, LVI and the number of LNs detected on PET-CT. In the multivariate model, LVI and detection of more than one LN on PET-CT were shown to be independent predictive values for detecting more than two ALNMs during SLNB or ALND (Table II).

To predict any number of LN metastases detected in SLNB or ALND, in the ROC analysis, the area under the curve at 4.05 cut-off value for pSUV_max 0.618 and the area under the curve at 1.25 cut-off value for nSUV_max 0.776 were calculated (Table III, Figure 1A). In the ROC analysis for the ability to predict more than two ALNMs in SLNB or ALND, the area under the curve at 4.65 cut-off value for pSUV_max 0.69 and the area under the curve at 1.75 cut-off value for nSUV_max 0.838 were calculated (Table III, Figure 1B).

DISCUSSION

Multidisciplinary protocols including surgery, radiation therapy, chemotherapy, targeted therapy and endocrine therapy are used in the treatment of breast cancer.⁹ Which combination of these treatments will be used is determined by the stage of the disease and its molecular subtype. Staging is based on tumour size, axillary lymph node involvement, and the presence of distant metastases. Axillary lymph node involvement alters both the treatment plan and the prognosis.¹⁰ In clinical practice, the axilla is routinely evaluated with USG, and if metastasis is suspected, a fine needle aspiration biopsy or trucut biopsy from the LN is performed.^11,12 Some studies reported that USG and PET-CT are nearly equal in axillary evaluation.^13,14 However, Davidson et al. utilised PET-CT to detect 18% of true positive lymph node metastases in patients with no suspicious findings in the axilla on USG.¹⁵ Similarly, Riegger et al. demonstrated that PET-CT was significantly more accurate than USG in detecting axillary breast cancer metastasis.¹⁶

In one of the earliest studies on the role of PET-CT in axillary staging, Veronesi et al. evaluated 236 clinical node-negative patients to compare PET-CT and SLNB. PET-CT findings were positive in 43 patients, with 38 being classified as true positives with pathological confirmation. However, 65 patients had axillary metastasis in SLNB. Many metastases could not be detected by PET-CT.¹⁷ In their study including 137 patients with early-stage breast cancer, Kim et al. reported the mean nSUV_max value in patients with ALNM to be higher than those without ALNM and found a cut-off value of ≥3.85 (sensitivity 50%, specificity 100%, and PPV 100%) for the presence of metastasis and <1.05 (sensitivity 100%, specific 33%, NPV 100%) for the absence of metastasis.¹⁸ Many other studies on the performance of PET-CT in detecting breast cancer-related axillary metastases have produced inconsistent results (sensitivity 24–82%, specificity 91–100%, PPV 63–100%, NPV 53–94% accuracy 78–94).^14,18-23 When this study and the other studies mentioned are considered together, the lack of a definite SUV_max cut-off value, as well as its low sensitivity and NPV, is the limiting factor for PET-CT axilla evaluation. The presence of involvement in PET-CT has a clear value, but the absence of involvement does not appear to imply that there is no metastasis in the axilla.

Today, it has been proven that axillary dissection is not required in every patient with SLNB metastasis. Many studies have shown that when patients are chosen based on the specific criteria, there is no significant difference in axillary recurrence and survival rates between those who received ALND and those who received direct or tangential radiotherapy to the axilla. In fact, less arm oedema was reported in the radiotherapy groups than in the ALND groups.^17,24 The ACOSOG Z0011 study is one of the most important studies that has changed the way medical professionals think about the axilla in recent years.⁸ The study included 891 patients who had breast-conserving surgery for a T1 or T2 invasive breast tumour, did not have palpable LNs in the axilla, and were on an adjuvant systemic therapy plan that included tangential whole-breast irradiation. All patients had one or two metastatic LNs in their SLNB. Only SLNB was performed in 446 patients and ALND in 445 patients. They revealed that there was no significant difference in 10-year overall survival or local recurrence rate and that the SLNB-only group had less arm morbidity.

The number of metastatic lymph nodes, as well as the presence of metastases in SLNB, have become critical in planning surgical treatment. The authors were unable to find another study in the literature that attempted to predict the number of metastatic LNs that can be found in SLNB using PET-CT. In this study, the nSUV_max cut-off value was calculated as 1.75 for the presence of more than two ALNM (sensitivity 81.8%, specificity 88.7%, PPV 67.6%, NPV 91.3%, accuracy 86.5%, p<0.001) (Table III, Figure 1B). In the univariate analysis, a relationship between pSUV_max, nSUV_max, LVI, and the number of LNs detected on PET-CT was discovered for the presence of more than two ALNMs, whereas in the multivariate analysis, a relationship between LVI and the number of LNs detected on PET-CT was discovered (Table II). In the study by Kim et al. detect the presence of any number of metastases, a very low nSUV_max cut-off value was calculated for the absence of metastases, while a high value was calculated for the presence of metastases. The metastasis status of patients who fall between 1.05 and 3.85 nSUV_max is uncertain.¹⁸ In the present study, for the presence of ALNM, the pSUV_max optimal cut-off value was 4.05, and the nSUV_max optimal cut-off value was 1.25 (Table III, Figure 1A). The cut-off value of 1.75, which the authors found for more than two ALNMs conditions, is exactly in the uncertain area in terms of metastasis in Kim et al.'s study. The authors think that this result is remarkable.

Table I: Patient characteristics.

		ALNM		Total	p-value
		Absence (n=42) n (%)	Presence (n=62) n (%)
Age*		58,7 ± 12,2	56,0 ± 11,5	57,1 ± 11,8	0,243
Histological grade	1	14 (33)	13 (21)	27 (26)
	2	22 (53)	34 (55)	56 (54)
	3	6 (14)	15 (24)	21 (20)	0,257
T-stage	1	15 (36)	10 (16)	25 (24)
	2	26 (62)	44 (71)	70 (67)
	3	1 (2)	6 (10)	7 (7)
	4	0 (0)	2 (3)	2 (2)	0,054
N-stage	0	40 (95)	0 (0)	40 (39)
	1	2 (5)	36 (58)	38 (36)
	2	0 (0)	18 (29)	18 (17)
	3	0 (0)	8 (13)	8 (8)	<0,001
LVI	Absence	35 (83)	24 (39)	59 (57)
	Presence	7 (17)	38 (61)	45 (43)	<0,001
PNI	Absence	31 (82)	33 (56)	64 (66)
	Presence	7 (18)	26 (44)	33 (34)	0,009
Ki-67	Low (<14%)	14 (33)	18 (29)	32 (30)
	High (≥14%)	28 (67)	44 (71)	72 (70)	0,641
ER	Negative	7 (17)	14 (23)	21 (20)
	Positive	35 (83)	48 (77)	83 (80)	0,461
PR	Negative	9 (21)	19 (31)	28 (27)
	Positive	33 (79)	43 (69)	76 (73)	0,298
HER-2	Negative	28 (67)	47 (76)	75 (72)
	Positive	14 (33)	15 (24)	29 (28)	0,308
Pathological type	Invasive breast carcinoma of no special type	30 (71)	55 (89)	85 (82)
	Invasive lobular carcinoma	5 (12)	2 (3)	7 (6)
	Cribriform carcinoma	3 (7)	2 (3)	5 (5)
	Invasive papillary carcinoma	3 (7)	2 (3)	5 (5)
	Neuroendocrine carcinoma	1 (3)	1 (2)	2 (2)	0,249
Number of LN detected on PET-CT	0	31 (74)	16 (26)	47 (45)
	1	6 (14)	18 (29)	24 (23)
	>1	5 (12)	28 (45)	33 (32)	<0,001
Operation method	BCS	31 (74)	22 (35)	53 (51)
	Mastectomy	11 (26)	40 (65)	51 (49)	<0,001
Type of axillary surgery	SLNB	32 (76)	3 (5)	35 (34)
	SLNB+ALND	4 (10)	24 (39)	28 (27)
	ALND	6 (14)	35 (56)	41 (39)	<0,001
More than 2 ALNM	No	42 (100)	29 (47)	71 (68)
	Yes	0 (0)	33 (53)	33 (32)	<0,001
*Mean Value ± Standard Deviation, ALNM: Axillary lymph node metastasis, LVI: Lymphovascular invasion, PNI: Perineural invasion, ER: Estrogen receptor, PR: Progesterone receptor, HER-2: Human epidermal growth factor receptor 2, LN: Lymph node, PET-CT: Positron emission tomography, BCS: Breast conserving surgery, SLNB: Sentinel lymph node biopsy, ALND: Axillary lymph node dissection.

Table II: Uni- and multi-variate logistic regression analysis for more than two axillary lymph node metastases.

		Univariate		p	Multivariate	p
		OR (%95 CI)			OR (%95 CI)
Age		1,005 (0,97 - 1,041)	0,791
pSUVmax		1,118 (1,004 - 1,246)	0,042
nSUVmax		1,535 (1,222 - 1,927)	<0,001
Histologic grade	1	Reference
	2	3,721 (1,132 - 12,224)	0,030
	3	2,875 (0,711 - 11,619)	0,138
LVI (Absence)		7,969 (3,084 - 20,591)	<0,001		5,783 (1,372 - 24,375)	0,017
PNI (Absence)		2,037 (0,84 - 4,939)	0,115
Ki-67 (<14)		1,278 (0,513 - 3,185)	0,599
ER (Negative)		0,542 (0,202 - 1,454)	0,224
PR (Negative)		0,412 (0,167 - 1,015)	0,054
HER-2 (Negative)		0,762 (0,296 - 1,961)	0,573
Number of LN Detected on PET-CT	0	Reference
	1	8,8 (2,102 - 36,847)	0,003		4,318 (0,677 - 27,541)	0,122
	>1	25,667 (6,537 - 100,784)	<0,001		17,544 (2,707 - 113,724)	0,003
Backward Wald method was used to include independent risk factors in the multivariate model. OR: Odds ratio, CI: Confidence interval, pSUVmax: Primary tumour maximum standardised uptake value, nSUVmax: Nodal maximum standardised uptake value, LVI: Lymphovascular invasion, PNI: Perineural invasion, ER: Estrogen receptor, PR: Progesterone receptor, HER-2: Human epidermal growth factor receptor 2, LN: Lymph node, PET-CT: Positron emission tomography.

Table III: ROC analysis results of pSUV_max and nSUV_max values for axillary lymph node metastasis presence.

		Cut-Off	AUC (%95CI)	p	Sensitivity (%95 CI)	Specificity (%95 CI)	PPV	NPV	Accuracy
ALNMa	pSUVmax	>4,05	0,618 (0,5 - 0,736)	0,042	0,581 (0,458 - 0,704)	0,548 (0,397 - 0,699)	0,672	0,470	0,568
	nSUVmax	>1,25	0,776 (0,686 - 0,867)	<0,001	0,613 (0,492 - 0,734)	0,929 (0,851 - 1,007)	0,700	0,619	0,741
>2 ALNMb	pSUVmax	>4,65	0,669 (0,565 - 0,773)	0,006	0,667 (0,506 - 0,828)	0,62 (0,507 - 0,733)	0,482	0,800	0,635
	nSUVmax	>1,75	0,838 (0,746 - 0,93)	<0,001	0,818 (0,686 - 0,95)	0,887 (0,813 - 0,961)	0,676	0,913	0,865
aPresence of any number of axillary lymph node metastases, b Presence of more than 2 axillary lymph node metastases, ROC: Receiver operating characteristic, AUC: Area under the curve, CI: Confidence interval, PPV: Positive predictive value, NPV: Negative predictive value, pSUVmax: Primary tumour maximum standardised uptake value, nSUVmax: Nodal maximum standardised uptake value.

The study method was designed to investigate the relationship between the nSUVmax value on PET-CT and the number of metastatic LNs in patients with breast cancer. The retrospective and single-centre design, the small number of patients, and the application of ALND to some patients and SLNB to others can be considered as the limitations of this study. On the other hand, since metastasis was not observed in the SLNB, ALND was not performed in 35 (34%) of the 104 patients included in the study. Sixty-nine (66%) patients underwent ALND because of at least one metastasis in the SLNB or because the SLN could not be identified. In total, more than two ALNMs were detected in 33 of the patients (32%). No axillary recurrence was observed in the follow-up of only SLNB patients and the ALND rate is relatively high. Therefore, the study method was optimal for pathological evaluation and the metastatic LNs count of the axilla.

The ACOSOG Z0011 study demonstrates the importance of the number of metastases, but it should be noted that only patients who underwent breast-conserving surgery and were clinical node-negative were included in this study. In patients with suspected lymph node metastasis in clinical evaluation, there is no evidence yet that the SUV_max value in PET-CT can be used instead of trucut biopsy or that it can be treated as the ACOSOG Z0011 study. The authors’ think that the relationship between the SUV_max value and the number of ALNM can take place in daily practice, supported by future randomised prospective studies. Targeting more than two ALNMs instead of at least one LN metastasis is seen to increase sensitivity and NPV and accuracy values that limit the use of PET-CT.

CONCLUSION

To avoid ALND in breast cancer patients, the number of metastases is as important as the presence of metastases in the SLNB. Low sensitivity, NPV, and accuracy values that limit the use of PET-CT in preoperative axillary evaluation can be increased by targeting the criterion of more than two ALNMs. Thus, PET-CT can be used more effectively in axilla management.

ETHICAL APPROVAL:
The study was approved by Sivas Cumhuriyet University Ethical Committee (Date: 21.09.2022 No. 2022-09/11).
 

PATIENTS' CONSENT:
Consent for the participation in the study was not obtained from patients as data were collected from medical record without disclosing the identity of participants.

COMPETING INTEREST:
The authors declared no competing interest.

AUTHORS’ CONTRIBUTION:
UK: Study conception and design.
UK, HT, GC, MCM: Acquisition of data.
UK, KK, ZH: Analysis and interpretation of data.
UK: Drafting of manuscript.
UK, KK, ZH: Critical revision.
All authors read and approved the final version of the manuscript to be published.

REFERENCES

1. Ferlay J, Shin HR, Bray F, Forman D, Mathers C, Parkin DM. Estimates of worldwide burden of cancer in 2008: GLOBOCAN 2008. Int J Cancer 2010; 127(12):2893-917. doi:10.1002/ijc.25516.
2. Groheux D. FDG-PET/CT for systemic staging of patients with newly diagnosed breast cancer. Eur J Nucl Med Mol Imaging 2017; 44(9):1417-9. doi: 10.1007/s00259-017- 3731-3.
3. Sawicki LM, Grueneisen J, Schaarschmidt BM, Buchbender C, Nagarajah J, Umutlu L, et al. Evaluation of ¹⁸F-FDG PET/MRI, ¹⁸F-FDG PET/CT, MRI, and CT in whole-body staging of recurrent breast cancer. Eur J Radiol 2016; 85(2):459-65. doi: 10.1016/j.ejrad.2015.12.010.
4. Valdora F, Houssami N, Rossi F, Calabrese M, Tagliafico AS. Rapid review: Radiomics and breast cancer. Breast Cancer Res Treat 2018; 169(2):217-29. doi: 10.1007/s10549-018- 4675-4.
5. Fatima N, Maseeh UZ. Role of ¹⁸FDG PET/CT in Breast Cancer. J Coll Physicians Surg Pak 2018; 28(3):177-9. doi: 10.29271/jcpsp.2018.03.177.
6. Kuru B. The adventure of axillary treatment in early stage breast cancer. Eur J Breast Health 2020; 16(1):1-15. doi: 10.5152/ejbh.2019.5157.
7. Donker M, van Tienhoven G, Straver ME, Meijnen P, van de Velde CJ, Mansel RE, et al. Radiotherapy or surgery of the axilla after a positive sentinel node in breast cancer (EORTC 10981-22023 AMAROS): A randomised, multi-centre, open-label, phase 3 non-inferiority trial. Lancet Oncol 2014; 15(12):1303-10. doi: 10.1016/S1470- 2045(14)70460-7.
8. Giuliano AE, Ballman KV, McCall L, Beitsch PD, Brennan MB, Kelemen PR, et al. Effect of axillary dissection vs. no axillary dissection on 10-year overall survival among women with invasive breast cancer and sentinel node metastasis: The ACOSOG Z0011 (alliance) randomised clinical trial. Jama 2017; 318(10):918-26. doi: 10.1001/ jama.2017.11470.
9. Groheux D, Hindie E. Breast cancer: Initial workup and staging with FDG PET/CT. Clinl Transl Imaging 2021; 9(3):221-31. doi: 10.1007/s40336-021-00426-z.
10. Iqbal J, Ginsburg O, Rochon PA, Sun P, Narod SA. Differences in breast cancer stage at diagnosis and cancer-specific survival by race and ethnicity in the United States. Jama 2015; 313(2):165-73. doi: 10.1001/jama.2014.17322.
11. Barco I, Chabrera C, García-Fernández A, Fraile M, González S, Canales L, et al. Role of axillary ultrasound, magnetic resonance imaging, and ultrasound-guided fine-needle aspiration biopsy in the preoperative triage of breast cancer patients. Clin Transl Oncol 2017; 19(6): 704-10. doi: 10.1007/s12094-016-1589-7. 
12. Tucker NS, Cyr AE, Ademuyiwa FO, Tabchy A, George K, Sharma PK, et al. Axillary ultrasound accurately excludes clinically significant lymph node disease in patients with early stage breast cancer. Ann Surg 2016; 264(6): 1098-102. doi: 10.1097/SLA.0000000000001549.
13. Kitajima K, Miyoshi Y. Present and future role of FDG-PET/CT imaging in the management of breast cancer. Jpn J Radiol 2016; 34(3):167-80.doi: 10.1007/s11604-015- 0516-0.
14. Ueda S, Tsuda H, Asakawa H, Omata J, Fukatsu K, Kondo N, et al. Utility of 18F-fluoro-deoxyglucose emission tomo-graphy/computed tomography fusion imaging (18F-FDG PET/CT) in combination with ultrasonography for axillary staging in primary breast cancer. BMC Cancer 2008; 8:165. doi: 10.1186/1471-2407-8-165.
15. Davidson T, Shehade N, Nissan E, Sklair-Levy M, Ben-Haim S, Barshack I, et al. PET/CT in breast cancer staging is useful for evaluation of axillary lymph node and distant metastases. Surg Oncol 2021; 38:101567. doi: 10.1016/ j.suronc.2021.101567.
16. Riegger C, Koeninger A, Hartung V, Otterbach F, Kimmig R, Forsting M, et al. Comparison of the diagnostic value of FDG-PET/CT and axillary ultrasound for the detection of lymph node metastases in breast cancer patients. Acta Radiol 2012; 53(10):1092-8. doi: 10.1258/ar.2012.110635.
17. Veronesi U, De Cicco C, Galimberti VE, Fernandez JR, Rotmensz N, Viale G, et al. A comparative study on the value of FDG-PET and sentinel node biopsy to identify occult axillary metastases. Ann Oncol 2007; 18(3):473-8. doi: 10.1093/annonc/mdl425.
18. Kim J, Lee J, Chang E, Kim S, Suh K, Sul J, et al. Selective sentinel node plus additional non-sentinel node biopsy based on an FDG-PET/CT scan in early breast cancer patients: Single institutional experience. World J Surg 2009; 33(5):943-9. doi: 10.1007/s00268-009-9955-z.
19. Heusner TA, Kuemmel S, Hahn S, Koeninger A, Otterbach F, Hamami ME, et al. Diagnostic value of full-dose FDG PET/CT for axillary lymph node staging in breast cancer patients. Eur J Nucl Med Mol Imag 2009; 36(10):1543-50. doi: 10. 1007/s00259-009-1145-6.
20. Koolen BB, Valdés Olmos RA, Elkhuizen PH, Vogel WV, Vrancken Peeters MJ, Rodenhuis S, et al. Locoregional lymph node involvement on 18F-FDG PET/CT in breast cancer patients scheduled for neoadjuvant chemotherapy. Breast Cancer Res Treat 2012; 135(1):231-40. doi: 10.1007/ s10549-012-2179-1.
21. Pritchard KI, Julian JA, Holloway CM, McCready D, Gulenchyn KY, George R, et al. Prospective study of 2-[¹⁸F] fluorodeoxyglucose positron emission tomography in the assessment of regional nodal spread of disease in patients with breast cancer: An Ontario clinical oncology group study. J Clin Oncol 2012; 30(12):1274-9. doi: 10200ucd20 11381103.
22. Taira N, Ohsumi S, Takabatake D, Hara F, Takashima S, Aogi K, et al. Determination of indication for sentinel lymph node biopsy in clinical node-negative breast cancer using preoperative 18F-fluorodeoxyglucose positron emission tomography/computed tomography fusion imaging. Jpn J Clin Oncol 2009; 39(1):16-21. doi: 10.1093/jjco/ hyn120. 
23. Zhang X, Wu F, Han P. The role of (18) F-FDG PET/CT in the diagnosis of breast cancer and lymph nodes metastases and micrometastases may be limited. Hell J Nucl Med 2014; 17(3):177-83.
24. Sávolt Á, Péley G, Polgár C, Udvarhelyi N, Rubovszky G, Kovács E, et al. Eight-year follow up result of the OTOASOR trial: The Optimal Treatment Of the Axilla - Surgery Or Radiotherapy after positive sentinel lymph node biopsy in early-stage breast cancer: A randomised, single centre, phase III, non-inferiority trial. Eur J Surg Oncol 2017; 43(4):672-9. doi: 10.1016/j.ejso.2016.12.011.