Preoperative immune checkpoint inhibition and cryoablation in early-stage breast cancer

Summary Local cryoablation can engender systemic immune activation/anticancer responses in tumors otherwise resistant to immune checkpoint blockade (ICB). We evaluated the safety/tolerability of preoperative cryoablation plus ipilimumab and nivolumab in 5 early-stage/resectable breast cancers. The primary endpoint was met when all 5 patients underwent standard-of-care primary breast surgery undelayedly. Three patients developed transient hyperthyroidism; one developed grade 4 liver toxicity (resolved with supportive management). We compared this strategy with cryoablation and/or ipilimumab. Dual ICB plus cryoablation induced higher expression of T cell activation markers and serum Th1 cytokines and reduced immunosuppressive serum CD4+PD-1hi T cells, improving effector-to-suppressor T cell ratio. After dual ICB and before cryoablation, T cell receptor sequencing of 4 patients showed increased T cell clonality. In this small subset of patients, we provide preliminary evidence that preoperative cryoablation plus ipilimumab and nivolumab is feasible, inducing systemic adaptive immune activation potentially more robust than cryoablation with/without ipilimumab.


Safety and tolerability
The primary endpoint of this study was reached, as all 5 women underwent standard-of-care primary breast surgery without delay.However, patients 1, 4, and 5 developed grade 1 transient hyperthyroidism.Patient 1 was lost to follow-up, but at the last toxicity visit, her free T4 was normal.In patient 4, hyperthyroidism resolved without further intervention.Patient 5 developed hyperthyroidism 3 weeks after surgery, leading to a delay in an unplanned axillary dissection at the surgeon's discretion.The patient ultimately decided not to proceed with the axillary dissection and her transient hyperthyroidism resolved without intervention after 6 weeks.Additionally, 8 weeks after primary breast surgery and 10 weeks after immunotherapy, patient 4 developed grade 4 liver toxicity.Prior to her diagnosis with breast cancer, she had a distant history of hepatic steatosis.At the time that she developed liver toxicity, she had been on exemestane for 4 weeks.She was admitted to the hospital and prescribed steroids and mycophenolic acid for 3 months.Liver biopsy showed no signs of immune-mediated hepatitis; this was complicated by the fact that she had received 24 h of steroids prior to biopsy.Pathologic review indicated that the injury pattern could be caused by drug-induced liver injury, including potentially from the exemestane, which was then discontinued.Five months later, once her liver function tests revealed amelioration of toxicity to grade 1, the patient was switched to tamoxifen.Initially, liver function tests performed over the first 3 months of treatment vacillated between normal and grade 1 elevations in aspartate aminotransferase (AST), alanine aminotransferase (ALT), and (alkaline phosphatase)ALP.Her most recent liver function tests were within the normal range.Repeat imaging of her liver is consistent with mild hepatic steatosis.AEs for all 5 patients are documented in Table 2.

Characterization of T cells in peripheral blood
Localized treatments inducing tumor destruction, such as cryoablation have been shown to induce immunogenic cell death that leads to systemic activation of the adaptive immune system. 15We investigated whether combining ipilimumab and nivolumab with cryoablation led to any changes in the frequencies of T cells or their activation status in peripheral blood using multicolor FACS analysis of peripheral blood mononuclear cells (PBMCs) banked from each time point outlined in Figure 1 and Table S1.We compared immune changes in the current patient cohort to those in patients from our original trial who were treated with ipilimumab or cryoablation alone or the combination of both. 10here were no substantial changes in the frequencies of CD3 + , CD8 + , or CD4 + Foxp3 -effector T cells in blood over baseline at any time after any treatment with ICB or cryoablation in this or the previous trial (Figure S1A).In general, cryoablation alone induced little to no changes in T cell frequencies or their activation in peripheral blood.In all patients treated with ipilimumab (alone or in combination with cryoablation), CD4 + Foxp3 + regulatory T cells became more abundant at 2 weeks (PC), but these cells' frequency started to decrease at 6 weeks (PS) postbaseline.Consistent with previous findings, 14 we found that 4PD-1 hi immunosuppressive cells expanded in patients treated with ipilimumab alone or in combination with cryoablation (Figures 2A and 2B).However, in the current cohort of patients receiving ipilimumab and nivolumab plus cryoablation, there was a decrease in the 4PD-1 hi T cell population at weeks 1 (PI) and 6 (PS), which led to a corresponding increase in the effector-to-suppressor cell ratios (Figures 2A and 2B).As the anti-PD-1 antibody used for FACS does not compete with nivolumab, any changes in PD-1 expression on T cells (e.g., CD4 + PD-1 hi ) cannot be attributed to an increase or decrease in the binding of the therapeutic antibody. 14ithin the time points examined, T cell activation peaked at 2 weeks post-baseline (PC) in all cohorts treated with ICB.In all cohorts, ICOS and Ki67 expression increased in both CD4 + and CD8 + T cells (Figure 2C).However, in the cohort receiving ipilimumab, nivolumab, and cryoablation Ki67 + PD-1 + CD4 + and CD8 + T cell populations increased.Ki67 + PD-1 + CD8 + T cells have been previously described as re-invigorated T cells in response to anti-PD-1 treatment. 16In summary, the triple combination of ipilimumab and nivolumab plus cryoablation appears to promote a robust CD4 + and CD8 + T cell activation profile peaking around 2 weeks post-treatment.
Few viable cells were recovered from cryopreserved single-cell suspensions of tumor collected at surgery, preventing reliable analysis of immune infiltrates and their activation status in the tumors by FACS (Figure S1B).
We currently do not have data from breast cancer patients treated with ipilimumab plus nivolumab alone for comparison and to assess the contribution of cryoablation in this triple combination.However, we have obtained and analyzed flow cytometry data on banked PBMCs from previously published work on patients with urothelial cancer (UC) and melanoma treated with ipilimumab plus nivolumab on similar timelines. 17In addition, these patients experienced mixed clinical responses within the subsets of UC (n = 10; 2 complete response [CR], 3 partial response [PR], 1 stable disease [SD], 3 progression of disease [PD]) and melanoma (n = 10[; 6 PR, 3 SD, 1 PD]).
In the UC and melanoma cohorts receiving ipilimumab plus nivolumab UC and melanoma, there was an initial increase in 4PD-1 hi cells which decreased in subsequent time points to below baseline, while the cryoablation plus ipilimumab plus nivolumab cohort showed a decrease in 4PD-1 hi cells at all time points (Figure S2A).This resulted in an increase in CD4 + and CD8 + to 4PD-1 hi ratios in the cryoablation plus ipilimumab plus nivolumab cohort at all time points and in the UC and melanoma cohorts at later time points (e.g., 6 weeks post-treatment [PS]).There was also an increase in the frequency of Tregs in the blood of ipilimumab + nivolumab cohorts which persisted over time; this was less pronounced in the patients with breast cancer receiving cryoablation plus ipilimumab plus nivolumab.Additionally, there was an overall increase in CD4 + and CD8 + T cell activation markers (e.g., ICOS, Ki67 and, to a lesser extent, CTLA-4, Lag-3 and, Tim-3) in the ipilimumab plus nivolumab cohorts, which peaked 1 week post treatment (PI) and decreased over time.The cryoablation plus ipilimumab plus nivolumab cohort of patients peaked at 2 weeks post treatment (PC) and then decreased over time.This suggests that the cryoablation therapy may influence T cell activation in this cohort (Figure S2B).There was a decrease in CD4 + and CD8 + T cell PD-1 expression in all cohorts over time.In summary, while there were some similarities in across all three cohorts, the addition of cryoablation plus ipilimumab plus nivolumab appeared to influence the magnitude and kinetics of T cell activation.However, we cannot exclude that these differences might be cancer specific (e.g., breast cancer vs. UC/melanoma).

Peripheral Th1 and Th2 cytokine response
We compared serum cytokine levels in the current study to those in patients treated with either ipilimumab or cryoablation alone, or the combination of both.In general, ipilimumab alone increased serum levels of certain Th2 cytokines such as IL-10 and IL-13, whereas the combination of ipilimumab and nivolumab plus cryoablation increased Th1 cytokines (Figure S3).The most substantial changes between these patient cohorts were seen in TNFa and IFNg.Serum TNFa increased in patients treated with either ipilimumab alone or cryoablation plus ipilimumab starting at 1 week (PI) and peaked at 6 weeks (PS) (Figure 2D), but this increase was not observed in patients treated with the combination of cryoablation plus ipilimumab and nivolumab.The combination of cryoablation plus ipilimumab and nivolumab substantially increased IFNg in the serum 6 weeks (PS).Thus, it appears that each of these treatment regimens may have differential effects on serum cytokine levels with distinct kinetics.
In addition to the Th1/Th2 cytokines analyzed previously, we also evaluated plasma from patients in the current study for additional cytokines and pro-inflammatory factors over time.In patients receiving ipilimumab and nivolumab plus cryoablation, levels of several pro-inflammatory factors such as c-reactive protein (CRP) and serum amyloid A (SAA) increased after cryoablation (2 weeks [PC]) (Figure 2E).Pro-inflammatory cytokines such as thymus and activation-regulated chemokine (TARC, CCL17) and monocyte chemoattractant protein-4 (MCP-4) also increased at weeks 1 (PI) and 6 (PS), respectively.

Characterization of the T cell repertoire in the tumor and periphery
We performed TCR sequencing in the 4 patients for whom samples were available to determine whether the combination of ipilimumab and nivolumab plus cryoablation alters the T cell repertoire in tumor or blood.In all 4 patients, the repertoire became more clonal following treatment with ipilimumab and nivolumab (Figure 3A, week 1 (PI) vs. 0 (Pre), as indicated by the relative size of the largest clones.This effect appeared to be transient, decreasing at later time points.
The dominance and kinetics of the top clones in the blood was mirrored by TCR clonality at these time points as measured by the Simpson index (Figure 3B).In all 4 patients, TCR clonality increased at week 1 (PI) and reverted to baseline at later time points.Thus, these changes in clonality can be attributed to treatment with ipilimumab and nivolumab but not cryoablation.On an individual basis, patients 1 and 2 had more diverse (less clonal) TCR repertoires while patients 3 and 4 had less diverse repertoires (more clonal).Patient 3 had a very dominating large clone at all time points.Patient 4 had an abnormally high clonal population of T cells that dominated the top 20% of T cells in the blood and had the largest increase in clonality following the treatment.We found substantial expansion of at least some T cell clones in blood at every time point in all patients (Figure 3C, red dots).Patient 1 had the most expanded T cell clones at 6 weeks, while patients 3 and 4 had the most expanded clones at 1 week (PI), indicating earlier response.Patient 2 had the fewest expanded clones, which may suggest less immune response.Tracking the frequency of clones that substantially expanded at 2 weeks (PC), likely reflecting response to cryoablation, over time revealed that these clones peaked at 2 weeks (PC) in patients 1 and 2, while they peaked at 1 week (PI) in patients 3 and 4 (Figure 3D).
Overall, the blood repertoire tells a different story for each patient, underscoring inter-person heterogeneity.Patient 1 had low clonality pre-treatment and robust clone expansion following the cryoablation.Patient 2 had similar clonality before treatment, but a less specific response.Patient 5 had one dominating clone and little change in other clones following treatment.Patient 4 also started with low clonality due to a few large clones, but displayed large expansion of these clones following treatment, reverting to the original state by week 6 (PS).
We also sequenced samples from the tumor (5 locations, collected PC) and surrounding tissues for TCR repertoire.In general, fewer productive TCRs were detected in tumor and surrounding tissue samples compared with blood and their frequency varied substantially among sites, making it difficult to draw conclusions (Table S2).We did not observe any shared trends in clonal frequencies, clonality, or expansion of clones among patients.Patient 1 had a dominant clone in most tissue samples, most abundant in the pre-treatment biopsy (Figure S4A), that did not appear in blood, indicating a tissue-resident clone.This patient also had the highest clonality score in the pre-treatment sample (Figure S4B).A few TCR clones in the core and surrounding tissues expanded PC (Figure S4C).
Morisita overlap index analysis, which is biased toward large clones, 18 revealed little overlap between clones present in tumor and normal breast tissue and blood (Figure S5).The overlap also varied on a patient-by-patient basis.Patients 1 and 3 showed an increase in overlap between blood and tumor after cryoablation, while patients 2 and 4 showed the opposite.Patient 1 showed the most overlap between Pre and PC core and tissue samples (Figure S5, green arrow).

DISCUSSION
In this pilot study, consistent with our primary aim, we demonstrated that the combination of ipilimumab, nivolumab, and cryoablation is feasible and can be integrated into preoperative treatment strategies without an AE delaying primary breast surgery.Cryoablation is believed to kill tumor cells via immunogenic cell death, releasing tumor-associated antigens that prime an immune response to the cancer, thereby serving as an in situ vaccine.While the implementation of immunotherapy has accelerated broadly across other cancer types, strategies that enhance immune responses and overcoming immune escape and acquired resistance remain elusive in breast cancer.For example, atezolizumab in combination with abraxane was originally granted accelerated approval for PD-L1-positive metastatic triple-negative breast cancer. 6More recently, however, atezolizumab was withdrawn from the market after a similar phase III study evaluating atezolizumab plus paclitaxel failed to meet the primary endpoint of progression-free superiority. 19At the same time, the demonstration that the addition of neoadjuvant pembrolizumab to chemotherapy for high-risk triple-negative breast cancer patients improves pathologic complete response rates and event-free survival 7 generates hope that amplifying the adaptive immune response to breast cancer is possible.
The varying effects of combining immunotherapy and chemotherapy underscore the need for novel techniques to enhance immunotherapies.We hypothesized that the addition of ipilimumab and nivolumab further enhances the immune response to the cancer locally and systemically.Neither this study of dual ICB nor our original study of single-agent ipilimumab and cryoablation were designed to assess clinical benefit; however, after 74 months median follow-up, there has been no evidence of metastatic disease among all 18 patients treated on the original study, including the 3 patients with TNBC.In 2015, at the time of the original trial design for the present study, the goal was to include women with HR+ disease based on the original aforementioned study effort evaluating cryoablation and ipilimumab across multiple subtypes (including HR+ disease).These original patients, including patients with HR+ disease, continued to have largely favorable long-term eventfree survival, and no concerning safety signals occurred.However, as the toxicity profiles were further elucidated, the nivolumab dose was modified to a flat dose in subsequent research trials, and patients with different cancer subtypes were included.Specifically, this work informed the ongoing phase 2 trial evaluating clinical benefit of ipilimumab, nivolumab, and cryoablation and survival outcomes exclusively in women with triple-negative breast cancer at high risk of recurrence.
With respect to strategies such as radiation versus cryoablation to enhance antigen presentation, the combination of immunotherapy with cryoablation (as opposed to radiation) in this study was chosen based on our earlier work and pre-clinical data available at the time.Specifically, pre-clinical and the aforementioned original clinical data indicated that cryoablation freezes macromolecules in their intact form and thereby preserves their antigenicity when thawed.Ionizing radiation causes degradation of these molecules so-at least theoretically-may  be less antigenic. 10Furthermore, the combination of ipilimumab and cryoablation recapitulated our original study demonstrating systematic clonal expansion in response to cryoablation induced antigen presentation. 12To date, there is no head-to-head data comparison of cryoablation versus radiation in enhancing antigen presentation.
Secondary aims of the study were to explore and characterize pre-and post-intervention peripheral blood and tumor responses to the combination of ipilimumab, nivolumab, and cryoablation.We compared our immune monitoring data to those of the previous study 10 and demonstrate stronger activation of T cells in the blood of patients treated with ipilimumab, nivolumab, and cryoablation compared with ipilimumab with or without cryoablation.In addition, we showed that the immune responses and kinetics in patients with breast cancer treated with cryoablation plus ipilimumab plus nivolumab were different from those of patients with UC or melanoma treated with ipilimumab plus nivolumab.This suggests that T cell activation is most pronounced when both ipilimumab and nivolumab are combined with cryoablation.The major caveat to this conclusion is that sample size is limited, and we did not evaluate nivolumab alone, nivolumab and ipilimumab, or cryoablation and nivolumab, which limits our understanding of the respective contribution of each agent.
In this study, we found dynamic changes in the population of 4PD-1 hi immunosuppressive T cells in the blood.The abundance of these cells decreased in patients treated with ipilimumab and nivolumab plus cryoablation, which increased the effector-to-suppressor cell ratio.In contrast, the size of this population notably increased in patients treated with treated with ipilimumab alone or in combination with cryoablation.This difference in effects on 4PD-1 hi activity between single-agent anti-CTLA-4 and the combination of anti-CTLA-4 with anti-PD-1 mirrors previously published data. 2 Whereas CTLA-4 blockade stimulates the accumulation of 4PD-1 hi cells in the tumor and periphery, PD-1 blockade opposes 4PD-1 hi activity even in combination with anti-CTLA-4, thereby promoting a stronger antitumor immune response. 2hile we found that the combination of ipilimumab and nivolumab plus cryoablation decreased the 4PD-1 hi T cell population at weeks 1 (PI) and 6 (PS), further studies are necessary to better characterize these effects over time, as well as their functional significance.The current findings suggest that the combination of ipilimumab and nivolumab engenders a uniquely potent immune response to cryoablation, given both the activation of effector T cells in the periphery and the mitigation of immunosuppressive 4PD-1 hi cells.
Patients receiving ipilimumab, nivolumab, and cryoablation showed increased levels of several pro-inflammatory factors in the blood at weeks 2 (PC) and 6 (PS).The release of pro-inflammatory Th1 cytokines has been strongly linked to antitumor immunity and is known to be promoted by immunotherapies. 20,21These cytokines are predominantly produced by activated T cells and have been linked with potent immune responses to both pathogens and cancer.While these cytokines can influence the activation of other immune cell types such as macrophages to enhance their effector function (e.g., phagocytosis), they can also directly inhibit tumor growth.IFNg has been strongly associated with anti-tumor efficacy in preclinical and clinical studies; 21 however, the role of TNFa in promoting tumor immunity is more ambiguous and most likely context dependent.Two studies examining serum TNFa levels after ICB treatment have drawn opposite conclusions in terms of clinical responses.In a study of non-small cell lung cancer (NSCLC), 22 patients treated with nivolumab or pembrolizumab showed an increase in serum levels of IFNg, TNFa, as well as many other pro-inflammatory cytokines, which correlated in better responses to anti-PD-1 inhibition and longer survival.However, in a study of melanoma, 23 there was a decrease in serum levels of TNFa in patients with CRs, PRs, and SD.In contrast, patients with PD showed an increase in levels of serum TNFa.In addition, there is preclinical evidence that combined ICB and TNFa blockade results in better anti-tumor responses and overall survival. 21Of note, in our study, we observed an increase in serum TNFa levels in patients treated with ipilimumab alone or cryoablation + ipilimumab at 6 weeks (PS) (Figure 2D) but this was not observed in the group receiving cryoablation alone or the group receiving cryoablation plus ipilimumab plus nivolumab.Furthermore, only the group receiving cryoablation plus ipilimumab plus nivolumab showed an increase in serum IFNg levels at weeks 1 (PI) and 2 (PC).Thus, TNFa and IFNg likely have different kinetics depending on treatment rendered.The cytokine data provided in this study is limited to 5-7 patients per treatment group; thus we cannot conclude whether the roles of these cytokines correlate positively or negatively with clinical responses.
TCR sequencing and diversity analysis in 4 patients revealed that combined ICB using ipilimumab and nivolumab led to an increase in T cell clonality in blood soon after treatment (pre-cryoablation) in all patients, followed by a decrease to pre-treatment values.Traditionally anti-CTLA-4 is known to diversify (render less clonal) the TCR repertoire, while anti-PD-1 is known to make the repertoire more clonal. 24This suggests that the combination of ipilimumab and nivolumab in this cohort behaves similarly to treatment with anti-PD-1.As with most effects of immunotherapies, the change in clonality was transient.The frequencies and expansion of TCR clones in the blood and tumor varied on a patient-by-patient basis, limiting further conclusions about these factors given that only 4 patients were analyzed in this cohort.
In conclusion, the original study of ipilimumab and cryoablation and the present study combining them with nivolumab fundamentally contribute to our evolving understanding of how immunotherapy can be used to improve outcomes for breast cancer patients.We have demonstrated enhanced immune responses with the addition of nivolumab to ipilimumab and cryoablation in the subset of patients studied.Essential to improving outcomes for patients will be the thoughtful integration of immunotherapy with a refined understanding of both shortand long-term AEs.Moreover, it will be critical to elucidate the relationship between the timing of immune changes and durable systemic responses in patients.As a result of this trial and as a next step toward evaluating immunotherapy and clinical benefit in breast cancer patients, our ongoing phase 2 study of perioperative ipilimumab, nivolumab, and cryoablation (NCT03546686) will evaluate event-free survival in a larger cohort of patients at especially high risk of recurrence.

Limitations of the study
There are several limitations to the work presented here.One major weakness is a small patient sample size which limits our ability to make broader inferences.In particular, this study is limited by the lack of diversity of the tumor subtypes given that all the included patients are ER+.All 5 patients who received dual ICB had ER+, HER2-breast cancer and 2 received adjuvant chemotherapy for axillary lymph node-positive disease.The heterogeneity in treatment and homogeneity of tumor type limit interpretations regarding biological and clinical outcome.Moreover, while the study met its primary endpoint of no delays of primary breast surgery, hyperthyroidism in one patient delayed a subsequently planned secondary axillary surgery.Further, patient 4 required a prolonged course of steroids in the setting of grade 4 liver toxicity, though this was not conclusively immune-related hepatitis, as she had already received several doses of steroids.As our understanding of immune-related toxicities and treatment for them evolves, it will be critical to identify patient populations for whom the toxicity risk versus potential benefit tips toward improving outcomes.To this end, our phase II trial includes only TNBC patients who do not achieve CR, and therefore have a substantially greater risk of distant recurrence and death.In this population, the potential benefits of such therapy may more significantly outweigh the risks of immune-related toxicities.To date, existing immunotherapies are more effective in triple negative breast cancer.Furthermore, we did not evaluate nivolumab alone, nivolumab and ipilimumab, or cryoablation and nivolumab, which limits our understanding of the respective contribution of each agent.Bariatrics, EndoWays, Impulse Dynamics, Innoblative Designs, Johnson & Johnson, Lantheus Medical Imaging PC, Motus GI Holdings Inc, Poseida Therapeutics Inc, and SureFire LLC; has provided professional services/activities to Advantagene Inc, Microbot Medical Ltd, Olympus (compensated) and XACT Robotics Ltd (uncompensated); and has equity, fiduciary role/position, and intellectual property rights in Aperture Medical Technology.L.N. reports: Agenus Inc, Celgene Cold Spring Harbor Laboratory, QLS Advisors LLC (professional services/activities); American Society of Clinical Oncology (ASCO), Breast Cancer Research Foundation, NewStem Ltd, Springer Nature Limited, Translational Breast Cancer Research Consortium, United States Department of Justice (professional services and activities, uncompensated); Martell Diagnostic Laboratories Inc. (equity); Codagenix Inc, Immix Biopharma Inc (equity; professional services/activities); Cure Breast Cancer Foundation (intellectual property rights; professional services/activities, uncompensated), Samus Therapeutics LLC (equity; fiduciary role/ position; professional services/activities, uncompensated).T.M. is a paid consultant for Immunos Therapeutics and Pfizer, is a co-founder and equity holder in IMVAQ Therapeutics, receives research support from Bristol Myers Squibb, Surface Oncology, Kyn Therapeutics, Infinity Pharmaceuticals, Inc., Peregrine Pharmaceuticals, Inc., Adaptive Biotechnologies, Leap Therapeutics, Inc., and Aprea, and is an inventor on patent applications related to work on oncolytic viral therapy, alpha virus-based vaccines, neoantigen modeling, CD40, GITR, OX40, PD-1, and CTLA-4.H.L.M. has consulted for Amgen, Bristol Myers Squibb, Celgene, Eli Lilly, Genentech/Roche, Immunomedics, Merck, OBI Pharma, Pfizer, Puma, Spectrum Pharmaceuticals, Syndax Pharmaceuticals, Peregrine, Calithera, Daiichi-Sankyo, Seattle Genetics, AstraZeneca, and TapImmune, and has received research support from Bristol Myers Squibb, MedImmune/AstraZeneca, BTG Ltd., and Merck.All reported research funding is administered by the institution.All disclosed relationships are outside the scope of the submitted work.The remaining authors declare no competing interests.

EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS
Between December 2016 and August 2017, women with biopsy-proven invasive breast cancer planning curative-intent mastectomy or lumpectomy at Memorial Sloan Kettering Cancer Center (MSK) were enrolled in this single-arm study (NCT02833233).Eligible women R 18 years of age, with R 1.5-cm, histologically confirmed, invasive carcinoma of the breast and no evidence of metastases were identified at the time of the initial surgical consultation and enrolled after providing informed consent.Male patients with early-stage BC were not included as BC primarily affects female patients, and inclusion of this population would result in a low sample size and render the study insufficiently powered to control for variables related to biological sex.Participants were enrolled regardless of race (breakdown: 60% white, 20% Asian, 20% unknown) or ethnicity (breakdown: 20% Hispanic or Latino/a of any race; 80% non-Hispanic/Latino/a).At the time of this clinical trial, our institution did not have a way to track socioeconomic indicators; we are building toward the capacity to track and report on those data with the rollout of an Epic-based electronic health record in February 2025.Any HR/HER2 and nodal status were permitted.HR positivity was defined as R 1% expression of either estrogen receptor (ER) or progesterone receptor by immunohistochemistry (IHC).HER2 positivity was defined as either 3+ expression by IHC and/or R 2.0 HER2 to chromosome 17 centromere signals by the FISH test.Multifocal, multicentric, and synchronous bilateral invasive disease was permitted.Patients were excluded from the study if they had a known autoimmune disease and/or were on immunosuppressants or steroids.Surgery and related management activities were scheduled per standard of care.Research study appointments and interventions did not interfere with planned standard-of-care surgery.All available imaging was reviewed to determine eligibility for magnetic resonance imaging (MRI)-or ultrasound (US)-guided cryoablation.Patients were advised to discontinue any non-steroidal antiinflammatory medications from the day of trial entry until 30 days after surgery, unless recommended by study investigator or required for the treatment of immune-related AEs.After providing informed consent, all 5 patients received ipilimumab and nivolumab 1-5 days prior to cryoablation, which was performed 7-10 days prior to primary breast surgery.At the time of cryoablation, at least three 9-or 12-gauge US-guided core biopsies of the primary breast tumor were obtained.If three core biopsies from each patient could not be obtained, patients were deemed ineligible.Two core breast specimens were provided to the MSK Department of Pathology for routine testing to confirm diagnosis and for immunohistochemical (IHC) staining.Additional core biopsy specimens were sent to the MSK Ludwig Center Immune Monitoring Core Facility for analyses reported herein.Patients then underwent safety assessments 2-3 weeks post-surgery (PS) and then every 2-3 weeks thereafter until at least 12 weeks post-immunotherapy (PI).Research blood samples were obtained at baseline, cryoablation, surgery, and at safety assessments after surgery (Figure 1).Tumor samples were obtained at cryoablation and surgery.This schedule mirrored our prior study of single agent ipilimumab with or without cryoablation. 10In the current study, ipilimumab was administered intravenously (IV) at a dose of 1 mg/kg infused over 90 minutes, combined with nivolumab at 3 mg/kg IV infused over 60 minutes.The 1 mg/kg ipilimumab dose was selected based on its better tolerability when combined with nivolumab. 25he study was performed in accordance with ethical principles of the Declaration of Helsinki and the International Conference on Harmonization of Good Practice and approved by the MSK Institutional Review Board.were run at deep resolution and all tissue samples were run at survey resolution using the Adaptive hsTCRBv4 assay.This Adaptive hsTCRBv4 platform has greater sensitivity than the version used in the previous ipilimumab plus cryoablation trial. 12

TCR analysis
TCR clones were defined from the Adaptive hsTCRBv4 assay based on their CDR3 amino acid sequence.A clone and count table were used for all further analysis.Repertoire diversity was assessed using the Simpson index, defined as the sum of the clone frequency squared divided by the frequency of all clones in a sample.This measure is equal to the probability of encountering the same clone twice while randomly sampling TCRs, and is higher for more clonal (less diverse) samples.Overlap between any two samples was measured using the Morisita similarity index, defined as the sum of the product of the clone frequencies from the two samples, normalized by the sum of the Simpson index of both samples. 18Thus, larger clones contribute more to this measure because differences in small clones are less substantial.The Morisita index is 0 for two samples that do not share any clone and 1 for identical samples.
Clones were also analyzed for substantial expansion between multiple samples from the same patient.A clone was deemed as expanded if its frequency increased by more than 2-fold and the Fisher exact test p-value was <0.05.This approach identifies clones that have expanded beyond what can be expected from sampling noise and largely filters out large fluctuation in small clones.

QUANTIFICATION AND STATISTICAL ANALYSIS
For flow cytometry and cytokine analyses where individual time points were compared, the p values were calculated using a paired t test.A p value of < 0.05 was considered statistically significant.For cytokine analysis where multiple treatment groups were compared, p values were calculated using 2-way ANOVA corrected for multiple comparisons.Results can be found in the Results section, as well as Figure 2A Representative bivariate plots of PD-1 vs. Foxp3 surface expression on CD3 + CD4 + T cells at baseline and 2 weeks post-treatment from a single patient from each treatment arm to identify CD4+PD-1 hi T cells [4PD-1 hi ]; B. Quantitation of CD4+PD-1 hi T cells and the ratio of CD8 + to 4PD- To restate, the statistics for TCR analysis, repertoire diversity was assessed using the Simpson index, defined as the sum of the clone frequency squared divided by the frequency of all clones in a sample.Overlap between any two samples was measured using the Morisita similarity index, defined as the sum of the product of the clone frequencies from the two samples, normalized by the sum of the Simpson index of both samples. 18Clones were also analyzed for substantial expansion between multiple samples from the same patient.A clone was deemed as expanded if its frequency increased by more than 2-fold and the Fisher exact test p-value was <0.05.Results can be found in the Results section, as well as Figure 3A

ADDITIONAL RESOURCES
Clinical trial registry number: NCT02833233.

Figure 2 .
Figure 2. The combination of ipilimumab, nivolumab, and cryoablation induces T cell activation in the periphery (A) Representative bivariate plots of PD-1 vs. Foxp3 surface expression on CD3 + CD4 + T cells at baseline and 2 weeks post-treatment from a single patient from each treatment arm to identify CD4 + PD-1 hi T cells (4PD-1 hi ).

Figure 3 .
Figure 3. T cell receptor (TCR) sequencing analysis of blood samples pre-and post-treatment (A) Frequencies of T cell clones ranked and color-coded, with the most abundant clones in color.(B) Simpson index for each time point.(C) Volcano plots of log 2 fold-change (Fc) versus -log 10 p value vs. pre-treatment.Lines through the y axis indicate a change in p value scale.(D) Frequency over time of the clones in the blood that expanded at 2 weeks compared with pre-treatment.
1 hi cells in each treatment arm.Cohort numbers are: cryoablation, n=7; ipilimumab, n=6; cryoablation plus ipilimumab, n=6; cryoablation plus ipilimumab plus nivolumab, n= 5; C. Heatmaps of expression of T cell activation markers in CD4 + T effector [Teff] cells and CD8 + T cells in each treatment arm.Data is represented as the average log fold-change (log FC) relative to baseline (t=0) for each time point; D. Comparison of TNFa and IFNg serum concentration between patients in each treatment arm.Statistics were calculated using 2-way ANOVA [*p<0.05,**p<0.01,***p<0.001,****p<0.0001].E. Serum concentrations of CRP, SAA, TARC, and MCP-4 pooled from 4 patients at baseline and 1, 2, and 6 weeks post-treatment with ipilimumab, nivolumab, and cryoablation.Statistics were calculated using paired t tests: *p<0.05,**p<0.01,***p<0.05), Figure S1A Heatmaps of T cell populations of the ratio of effector to suppressor T cells in each treatment group.Cohort numbers are: cryoablation, n=7; ipilimumab, n=6; cryoablation plus ipilimumab, n=6; cryoablation plus ipilimumab plus nivolumab, n=5; B) Banked single-cell suspensions of tumor-infiltrating lymphocytes isolated from the tumors were analyzed by flow cytometry.Shown are the frequencies of total CD3 + , CD8 + ,-and CD4 + effector [CD4+Foxp3-] T cells, Tregs [CD4+Foxp3+], and 4PD1 hi [CD4+Foxp3-PD1 hi ] T cells at various tumor and tissue sites +/-standard error of samples pooled from 3-5 patients.Normal S/I, normal tissue pooled from superior and inferior samples.Statistics were calculated using Student's t test: *p<0.05,**p<0.01,***p<0.05), Figure S2 (Flow cytometry data was obtained from published work 17 on patients with UC [n=94] and melanoma [n=188] treated with ipilimumab plus nivolumab and a similar analysis was performed as in the cohort of patients with breast cancer cohort receiving cryoablation plus ipilimumab plus nivolumab: A) Heatmaps of T cell populations of the ratio of effector to suppressor T cells in each treatment group.Data is represented as the average log10 fold-change [log FC] relative to baseline [t=0, Pre] for each time point; B) Heatmaps of expression of T cell activation markers in CD4 + T effector [Teff] cells and CD8 + T cells in each cohort), and S3 (Heatmaps of serum Th1 and Th2 cytokines in serum in each treatment group.Data for each time point are represented as the average of log 10 fold change [FC] relative to the baseline levels (pg/ml) for each cytokine.Cohort numbers are: cryoablation, n=7; ipilimumab, n=6; cryoablation plus ipilimumab, n= 6; cryoablation plus ipilimumab plus nivolumab, n=5).
Frequencies of T cell clones ranked and color-coded; B. Simpson index for each time point; C. Volcano plots of log 2 fold-change [Fc] versus -log 10 p-value vs. pre-treatment; D. Frequency over time of the clones in the blood that expanded at 2 weeks compared with pre-treatment; n=4), Figure S4A Frequencies of T cell clones in the tumor ranked by abundance; B. Simpson index for each time point in the tumor; C. Volcano plots of log 2 fold change (Fc) vs. negative log 10 p-value of data compared with pre-treatment in blood), Figure S5 (Heatmaps of the calculated Morisita overlap index between blood and tissue; n=4), and Table S2 (Number of productive T cell receptors [TCRs] per patient sample; n=4).

Immunotherapy Cryo Breast surgery Blood draw 6-8 days 6-10 days 29 37 days Core biopsy + Blood draw Blood draw Blood draw -
Figure 1.Study schemaCombined immune checkpoint blockade (immunotherapy) was administered 1-5 days prior to, and cryoablation (Cryo) was performed 7-10 days prior to standard-of-care surgery.Toxicity evaluation continued for 12 weeks after drug administration.Blood for immune correlates was obtained at baseline, cryoablation, surgery, and 2-4 weeks thereafter (see TableS1for individual patient timelines).Tumor samples were obtained at cryoablation and surgery.