Table of Contents
- Key Points
- Background: Why This Question Matters
- How the Study Was Conducted
- Who Took Part in the Study
- What Treatments Patients Received
- What the Study Measured
- How the Statistics Worked
- Key Findings: Survival Results
- Recurrence and Death Rates in Detail
- Safety and Side Effects
- What This Means for Patients
- Limitations of the Study
- Recommendations and Next Steps
- Frequently Asked Questions
- Source Information
Key Points
- In the INSEMA trial of 5,502 women with early-stage, node-negative breast cancer, skipping axillary surgery was noninferior to sentinel-lymph-node biopsy for 5-year invasive disease-free survival.
- Estimated 5-year invasive disease-free survival was 91.9% with surgery omission and 91.7% with sentinel-node biopsy; the hazard ratio was 0.91 (95% CI, 0.73 to 1.14).
- Patients who skipped axillary surgery had less lymphedema, better arm mobility, and less arm or shoulder pain than those who had sentinel-node biopsy.
- Axillary recurrence occurred in 1.0% of the surgery-omission group versus 0.3% of the sentinel-node biopsy group; death occurred in 1.4% versus 2.4%, respectively.
- Most patients in the trial had small tumors (90% clinical T1, 79% pathological T1); patients with larger tumors should discuss whether these findings apply to them.
Background: Why This Question Matters
For many years, doctors considered the status of the lymph nodes under the arm (the axillary nodes) to be one of the most important prognostic factors in invasive breast cancer. Along with tumor size, lymph node status helped guide decisions about systemic therapy (drug treatment that travels through the body) and radiotherapy (radiation treatment).
More recently, researchers have recognized that the molecular characteristics of a tumor also matter for how a patient responds to therapy and how long she survives. This means that decisions about systemic therapy must now balance both nodal status and tumor biology.
The publication of results from the American College of Surgeons Oncology Group (ACOSOG) Z0011 trial provided the basis for the design and funding of several de-escalation trials. De-escalation means reducing the intensity of treatment when it is safe to do so.
Four prospective, randomized trials are investigating whether axillary sentinel-lymph-node biopsy can be omitted in patients with clinically node-negative (cN0) breast cancer who undergo upfront breast-conserving surgery. These trials are:
- SOUND (Sentinel Node versus Observation after Axillary Ultrasound)
- INSEMA (Intergroup Sentinel Mamma)
- BOOG 2013-08 (Dutch Breast Cancer Research Group 2013-08)
- NAUTILUS (No Axillary Surgical Treatment for Lymph Node-Negative Patients after Ultra-Sonography)
Recently, primary results from the SOUND trial showed that omission of axillary surgery was noninferior to sentinel-lymph-node biopsy in patients with small breast cancers up to 2 cm. The primary outcome of the SOUND trial was distant disease-free survival at 5 years, analyzed in the intention-to-treat population of 1,405 patients after a median follow-up of 5.7 years.
Primary-outcome results of the BOOG trial, which involved 1,644 patients, are expected to be reported in 2025. Later, in 2027, data from the NAUTILUS trial, which involved 1,734 patients, are expected.
The INSEMA trial also included a preplanned central quality-assurance review process for radiotherapy planning and axillary contouring. The findings of that review were published in 2020. A recent analysis of patient-reported outcomes showed differences between the groups in quality-of-life scores for breast symptoms and arm symptoms. Patients who did not undergo sentinel-lymph-node biopsy had better scores than those who did.
In the assessment of arm symptoms, patients who underwent sentinel-lymph-node biopsy had significantly higher scores, indicating worse symptoms, for pain, arm swelling, and impaired mobility at all postoperative visits.
The goal of the INSEMA trial is to show that complete omission of axillary surgery in early-stage breast cancer treated with breast-conserving surgery is noninferior to sentinel-lymph-node biopsy. The goal is noninferiority with respect to invasive disease-free survival.
How the Study Was Conducted
The INSEMA trial was a prospective, randomized, noninferiority trial conducted at 142 sites in Germany and 9 sites in Austria. All sites obtained approval from local independent review boards before the trial began.
The trial was designed by the University Medicine Rostock in cooperation with the German Breast Group. The German Breast Group performed data management, statistical analysis, project management, monitoring, and trial oversight. The Austrian Breast and Colorectal Cancer Study Group provided regulatory coordination at trial sites in Austria.
All verification of source data was performed according to the standard operating procedures of the German Breast Group. The first author and the statistician vouch for the completeness and accuracy of the data and for the fidelity of the trial to the protocol.
The first draft of the manuscript was written by the first author in collaboration with the German Breast Group. No one who is not an author contributed to the writing of the manuscript. All authors contributed substantially and approved the version of the manuscript that was submitted for publication.
Who Took Part in the Study
Women with breast cancer who planned to undergo upfront breast-conserving surgery were eligible for the trial. To qualify, patients had to be at least 18 years of age and have a clinical tumor stage of T1 or T2, meaning a tumor size of 5 cm or less. They also had to have node-negative status according to clinical assessment (cN0) and imaging (iN0). All patients provided written informed consent.
Patients were first randomly assigned in a 1:4 ratio. This means that for every one patient assigned to skip axillary surgery, four patients were assigned to undergo sentinel-lymph-node biopsy. The surgery-omission group received treatment without axillary surgery. The surgery group underwent sentinel-lymph-node biopsy.
Patients who underwent sentinel-lymph-node biopsy and were found to have pathological sentinel-node-positive status with one to three macrometastases (cancer deposits larger than 2 mm) underwent a second randomization. In that second randomization, they were assigned in a 1:1 ratio to undergo completion axillary-lymph-node dissection or to proceed without dissection, having undergone sentinel-lymph-node biopsy alone. The results of the second randomization are not reported in this article.
During follow-up, patients were assessed according to standard clinical practice. A medical history was taken and a physical examination was performed every 6 months for the first 36 months and yearly thereafter. Annual mammography and sonography were required. Other testing was performed according to the patient's symptoms at the discretion of the investigator.
What Treatments Patients Received
The preoperative diagnostic workup included routine axillary ultrasonography (ultrasound imaging of the underarm area) before biopsy. In cases of cN0 status, ultrasonography indicated the presence of cancer in a lymph node (iN+ status). In these cases, a negative core biopsy of the lymph node was required before randomization. A negative fine-needle aspiration biopsy of the lymph node was also required before randomization.
For identifying metastatic lymph nodes by axillary ultrasonography, the recommended criteria were a cortical thickness greater than 2.5 mm or the absence of a fatty hilum.
All patients underwent unilateral breast-conserving surgery with postoperative whole-breast irradiation, regardless of the intrinsic cancer subtype. Conventional fractionation or moderate hypofractionation could be used. Radiation therapy could be delivered with three-dimensional conformal or intensity-modulated techniques.
The axilla was not specifically targeted with radiation. Use of high tangents or regional nodal irradiation was not permitted, except for patients with four or more axillary lymph node metastases in the surgery group.
A boost of radiotherapy to the tumor bed was generally recommended. However, the boost could be omitted in patients at lower risk for local recurrence. These were patients older than 60 years of age with a small tumor size and favorable tumor biology. The use of partial breast irradiation was not allowed.
Procedures to ensure the quality of radiotherapy included a central review of radiotherapy plans for the first three patients treated at each center. Dosimetric data were collected in the electronic case report forms for the entire trial population.
What the Study Measured
The primary outcome for the first randomization was invasive disease-free survival. This was prespecified as the period between randomization and the first primary-outcome event. A primary-outcome event was defined as one of the following:
- Recurrence of local, axillary, or distant invasive disease
- Death from any cause
- Occurrence of contralateral invasive breast cancer (cancer in the other breast)
- Occurrence of a second primary invasive cancer of a type other than breast cancer
Because fewer patients had positive sentinel-lymph-node biopsies than expected, invasive disease-free survival was downgraded from a primary to a secondary outcome for the second randomization. This change was made according to protocol amendment 5. The second randomization assigned patients to undergo sentinel-lymph-node biopsy alone or completion axillary-lymph-node dissection.
Other secondary outcomes included overall survival, locoregional disease-free survival, and ipsilateral axillary recurrence (cancer returning in the same underarm area). Other secondary outcomes also included distant disease-free survival, quality-of-life measures, and dose distribution in ipsilateral axilla levels I through III during radiotherapy.
Patient-reported outcomes were assessed with the European Organisation for Research and Treatment of Cancer (EORTC) 30-item Quality of Life of Cancer Patients questionnaire QLQ-C30. Patient-reported outcomes were also assessed with the EORTC Breast Cancer questionnaire QLQ-BR23. These assessments were done at baseline before surgery and at 1, 3, 6, 12, and 18 months after surgery.
Scores for all the multi-item scales and single-item measures range from 0 to 100. Higher scores on measures of global health status and quality of life indicate a higher quality of life. However, higher scores on multi-item symptom scales or individual symptom-related items indicate worse symptoms.
Short-term surgical complications within 4 weeks after the final surgery were recorded. Selected long-term complications were documented during the complete follow-up period.
How the Statistics Worked
To calculate the sample size, the researchers considered the 5-year invasive disease-free survival rate among patients with cN0 or iN0 and T1 or T2 disease to be 88%.
To show clinical noninferiority of the omission of axillary surgery, the researchers specified two requirements. First, the surgery-omission group would need to have a 5-year invasive disease-free survival rate of at least 85%. Second, the upper end of the 95% confidence interval for the hazard ratio for invasive disease or death in the surgery-omission group would need to be less than 1.271. This hazard ratio was compared with the surgery group.
The overall error rate for a false-positive outcome (alpha) was set to 5%. The error rate for a false-negative result (beta) was set to 20%. This means the trial was calculated to have 80% power to confirm noninferiority according to the specified criteria.
After adjustment for randomization in a 1:4 ratio according to the method of Hsieh, the researchers calculated that a total of 851 primary-outcome events would be needed for the first randomization. The researchers also calculated that a per-protocol population of 5,230 patients would be needed for the first randomization. This per-protocol population would include 1,046 patients in the surgery-omission group and 4,184 patients in the surgery group.
With the assumption that 5% of patients would have to be excluded from the per-protocol analysis, approximately 5,505 patients would be needed for the first randomization. The researchers planned for the final efficacy analysis to be event driven. It would be performed when 851 primary-outcome events for the first randomization had occurred in the per-protocol population.
In the case of lower rates of events than expected, the researchers planned for a time-driven analysis to be performed after the last-recruited patient had completed a follow-up of 5.5 years.
The analysis of the primary outcome was performed with the per-protocol population because of the noninferiority design of the trial. Primary-outcome results for the per-protocol and intention-to-treat populations are reported. The results of a sensitivity analysis that does not exclude patients who did not undergo radiotherapy are also reported.
The analyses were performed with data available as of August 30, 2024, after nearly 5.5 years of follow-up for the last patient enrolled, because 851 events had not yet accrued. The Kaplan-Meier product-limit method was used to estimate rates of 5-year invasive disease-free survival, reported with two-sided 95% confidence intervals.
Noninferiority was tested on the basis of the confidence interval of the hazard ratio from the Cox proportional-hazards model. The goal was to exclude a hazard ratio of 1.271 or above for invasive disease or death. A multivariate Cox proportional-hazards model was used to adjust hazard ratios according to stratification factors: age, tumor size, and tumor grade.
The homogeneity of findings among subgroups stratified according to age, tumor size, tumor grade, and histologic subtype was explored with the use of univariate Cox regressions. The Pocock minimization method was used for trial-group assignment, with groups stratified according to specified stratification criteria.
Analyses were performed with SAS, version 9.4, with SAS Enterprise Guide version 8.3. No adjustment for multiple comparisons was made. All confidence intervals and tests were two-sided. The widths of the confidence intervals have not been adjusted for multiple comparisons and should not be used in place of hypothesis testing of secondary outcomes.
Key Findings: Survival Results
Between September 2015 and April 2019, a total of 5,502 patients were recruited for the first randomization. Of these, 5,154 were included in the intention-to-treat population: 1,030 assigned to undergo treatment without sentinel-lymph-node biopsy and 4,124 assigned to undergo sentinel-lymph-node biopsy.
A total of 348 patients (6.3%) withdrew from the trial. The main reasons for withdrawal were secondary mastectomy and withdrawal of consent before completion of breast surgery.
After exclusion of an additional 296 patients for various reasons, such as lack of postoperative radiotherapy, a total of 4,858 patients were included in the per-protocol population. Of these, 962 were assigned to undergo treatment without sentinel-lymph-node biopsy and 3,896 to undergo sentinel-lymph-node biopsy.
All baseline characteristics were well balanced between the trial groups. The demographic and clinical characteristics of the trial population were similar to those of the general population of patients who undergo treatment for breast cancer in Germany and Austria.
Data on race or ethnic group were not collected specifically as part of the trial. However, the majority of the patients were White, which reflects the general population in the areas surrounding the trial sites. This may not be representative of patients with breast cancer in other geographic regions.
The median age at cancer diagnosis was 62.0 years, with a range of 24.0 to 89.0 years and an interquartile range of 53.0 to 68.0 years. Only 527 patients (10.8%) were enrolled at less than 50 years of age.
The median preoperative tumor size was 15 mm (interquartile range, 10 to 20) as assessed by palpation and 11 mm (interquartile range, 8 to 16) as assessed by imaging. Of the patients who underwent imaging, 94.7% underwent sonography.
A total of 4,392 patients (90.4%) had cancer staged as clinical T1, and 3,855 patients (79.4%) had pathological T1 cancer. Among the subgroup of 466 patients (9.6%) with clinical T2 cancer, the median tumor size in both trial groups was 25 mm. The interquartile range was 22 to 29 in the surgery-omission group and 22 to 28 in the surgery group.
In the surgery group, 579 of 3,854 patients (15.0%) with sentinel lymph nodes detected and a known nodal status had cancer-positive sentinel nodes. Of these, 133 patients (3.5%) had micrometastases, 438 (11.4%) had one to three macrometastases, and 8 (0.2%) had at least four positive nodes.
The incidence of sentinel-node positivity for micro- and macrometastases was higher in the subgroup of patients with T2 cancer than in the subgroup with T1 cancer.
The median follow-up was 73.6 months. The estimated 5-year invasive disease-free survival rate was 91.9% (95% confidence interval [CI], 89.9 to 93.5) among patients in the surgery-omission group. The rate was 91.7% (95% CI, 90.8 to 92.6) among patients in the surgery group. The hazard ratio was 0.91 (95% CI, 0.73 to 1.14), which was below the prespecified noninferiority margin.
Recurrence and Death Rates in Detail
The analysis of the first primary-outcome events included the occurrence or recurrence of invasive disease or death from any cause. These events occurred in a total of 525 patients (10.8%).
The analysis showed apparent differences between the surgery-omission group and the surgery group in the incidence of axillary recurrence. Axillary recurrence occurred in 1.0% of patients in the surgery-omission group and 0.3% of patients in the surgery group.
The analysis also showed apparent differences in the incidence of death. Death occurred in 1.4% of patients in the surgery-omission group and 2.4% of patients in the surgery group.
These findings come from the per-protocol analysis, which was the prespecified primary analysis for this noninferiority trial.
Safety and Side Effects
The safety analysis indicates that patients in the surgery-omission group had a lower incidence of lymphedema than patients who underwent sentinel-lymph-node biopsy. Lymphedema is swelling caused by fluid buildup, usually in the arm.
Patients in the surgery-omission group also had greater arm mobility than patients who underwent sentinel-lymph-node biopsy.
In addition, patients in the surgery-omission group had less pain with movement of the arm or shoulder than patients who underwent sentinel-lymph-node biopsy.
These safety findings are consistent with the earlier analysis of patient-reported outcomes. That analysis showed that patients who underwent sentinel-lymph-node biopsy had significantly higher scores, indicating worse symptoms, for pain, arm swelling, and impaired mobility at all postoperative visits.
What This Means for Patients
In this trial involving patients with clinically node-negative, T1 or T2 invasive breast cancer, omission of surgical axillary staging was noninferior to sentinel-lymph-node biopsy. This finding held after a median follow-up of 6 years.
Most patients in the trial had small tumors. Specifically, 90% had clinical T1 cancer and 79% had pathological T1 cancer.
For patients who fit this profile, the results suggest that skipping underarm surgery does not compromise their chances of remaining free of invasive disease. The 5-year invasive disease-free survival rates were nearly identical between the two groups: 91.9% with surgery omission and 91.7% with sentinel-node biopsy.
The hazard ratio of 0.91 means that, after accounting for other factors, the risk of invasive disease or death was slightly lower in the surgery-omission group. The confidence interval of 0.73 to 1.14 was entirely below the noninferiority margin of 1.271, which is why the trial met its goal.
At the same time, patients who skipped axillary surgery experienced fewer side effects. They had less lymphedema, better arm mobility, and less arm or shoulder pain. This is an important quality-of-life benefit, because arm problems are a common long-term concern after breast cancer surgery.
The trial did show a small difference in axillary recurrence: 1.0% in the surgery-omission group versus 0.3% in the surgery group. This difference was apparent but small. The trial also showed a difference in death rates: 1.4% in the surgery-omission group versus 2.4% in the surgery group.
Limitations of the Study
The trial did not collect data on race or ethnic group. The majority of patients were White, which reflects the general population in the areas surrounding the trial sites. The results may not be representative of patients with breast cancer in other geographic regions.
The analysis was performed with data available as of August 30, 2024, because 851 events had not yet accrued. The researchers therefore performed a time-driven analysis after nearly 5.5 years of follow-up for the last patient enrolled, rather than an event-driven analysis.
The primary outcome was analyzed in the per-protocol population rather than the intention-to-treat population. This approach was chosen because of the noninferiority design of the trial. Results for both populations are reported.
No adjustment for multiple comparisons was made. The widths of the confidence intervals have not been adjusted for multiple comparisons and should not be used in place of hypothesis testing of secondary outcomes.
The results of the second randomization, which involved patients with one to three positive sentinel nodes, are not reported in this article.
Recommendations and Next Steps
Patients with early-stage, clinically node-negative breast cancer who are scheduled for breast-conserving surgery may wish to discuss the option of omitting axillary surgery with their treatment team. The INSEMA trial suggests this de-escalation concept may be suitable for patients 50 years of age or older with low-risk (grade G1 or G2), HR-positive, HER2-negative invasive breast cancer and clinical T1 tumors.
It is important to note that most patients in this trial had small tumors. Specifically, 90% had clinical T1 cancer and 79% had pathological T1 cancer. Patients with larger tumors should discuss whether these findings apply to their situation.
Patients should also be aware that the trial found a small difference in axillary recurrence between the two groups. The rate was 1.0% in the surgery-omission group and 0.3% in the surgery group. This difference should be weighed against the benefits of less lymphedema, better arm mobility, and less pain.
The INSEMA trial is one of several de-escalation trials. Results from the BOOG trial, involving 1,644 patients, are expected in 2025. Results from the NAUTILUS trial, involving 1,734 patients, are expected in 2027. These trials will provide additional evidence about whether axillary surgery can be safely omitted.
The SOUND trial previously showed that omission of axillary surgery was noninferior to sentinel-lymph-node biopsy in patients with small breast cancers up to 2 cm. Together, these trials are changing how doctors think about axillary surgery in early breast cancer.
Frequently Asked Questions
What was the INSEMA trial testing?
The INSEMA trial tested whether women with early-stage, clinically node-negative breast cancer could safely skip underarm (axillary) surgery. It compared skipping axillary surgery with sentinel-lymph-node biopsy in 5,502 women who had breast-conserving surgery. The goal was to see if omitting surgery was not worse than biopsy for invasive disease-free survival.
Who was eligible for the INSEMA trial?
Women were eligible if they were at least 18 years old, had a tumor 5 cm or smaller (stage T1 or T2), and had no evidence of cancer in the underarm lymph nodes by clinical assessment and imaging. They also had to plan to undergo upfront breast-conserving surgery. All patients provided written informed consent.
What did the trial find about survival?
After about six years, the estimated 5-year invasive disease-free survival rate was 91.9% in the surgery-omission group and 91.7% in the sentinel-node biopsy group. The hazard ratio was 0.91, meeting the trial's noninferiority criteria. This means skipping axillary surgery was not worse than sentinel-node biopsy for this outcome.
What were the side effects of skipping axillary surgery?
Patients who skipped axillary surgery had less lymphedema (arm swelling), better arm mobility, and less arm or shoulder pain compared with those who had sentinel-lymph-node biopsy. These benefits were seen in patient-reported outcomes and safety analyses. The trial showed that omitting surgery reduced these common long-term arm problems.
Was there any difference in cancer recurrence in the underarm area?
Yes, a small difference was seen. Axillary recurrence occurred in 1.0% of patients who skipped axillary surgery and 0.3% of those who had sentinel-node biopsy. This difference was apparent but small. Patients should weigh this against the benefits of less lymphedema, better arm mobility, and less pain.
What does a hazard ratio of 0.91 mean?
A hazard ratio of 0.91 means that, after accounting for other factors, the risk of invasive disease or death was slightly lower in the surgery-omission group compared with the sentinel-node biopsy group. The confidence interval (0.73 to 1.14) was entirely below the noninferiority margin of 1.271, so the trial met its goal.
What are the limitations of the INSEMA trial?
The trial did not collect data on race or ethnic group, and most patients were White, so results may not represent all geographic regions. The analysis was time-driven because 851 events had not accrued. The primary outcome was analyzed in the per-protocol population, and no adjustment was made for multiple comparisons.
I have early-stage, node-negative breast cancer and my surgeon wants to skip underarm lymph node surgery — when should I get a second opinion?
A second opinion is reasonable when you are deciding whether to omit axillary surgery before breast-conserving surgery. In a trial of 5,502 women with clinically node-negative T1 or T2 tumors, 5-year invasive disease-free survival was 91.9% without axillary surgery versus 91.7% with sentinel-node biopsy, meeting noninferiority. Skipping surgery meant less lymphedema, better arm mobility, and less arm or shoulder pain, though axillary recurrence was 1.0% versus 0.3%. Most participants had small tumors, so patients with larger tumors should confirm how these findings apply. Diagnostic Detectives Network provides independent expert second opinions.
Source Information
Original article title: Axillary Surgery in Breast Cancer Primary Results of the INSEMA Trial NEJM 2024
Authors: T. Reimer, A. Stachs, K. Veselinovic, T. Kühn, J. Heil, S. Polata, F. Marmé, T. Müller, G. Hildebrandt, D. Krug, B. Ataseven, R. Reitsamer, S. Ruth, C. Denkert, I. Bekes, D.-M. Zahm, M. Thill, M. Golatta, J. Holtschmidt, M. Knauer, V. Nekljudova, S. Loibl, and B. Gerber
Publication: The New England Journal of Medicine, March 13/20, 2025, volume 392, number 11, pages 1051-1064
DOI: 10.1056/NEJMoa2412063
Funding: German Cancer Aid
ClinicalTrials.gov number: NCT02466737
This patient-friendly article is based on peer-reviewed research.
