With the increasing use of ICI in the treatment of breast cancer, early signs of treatment benefit as well as predictors of immune-related toxicity are needed in order to select patients for whom the benefits of immunotherapy outweigh the risks. Early radiologic markers of response represent one possible predictive tool, particularly with the use of serial breast MRIs, which are already commonly used in the setting of neoadjuvant therapy.(9, 11, 13) Prior serial breast MRI evaluation has focused on changes with neoadjuvant chemotherapy alone, with decreasing tumor size and FTV predictive of response.(9, 13, 16, 17, 19) However, in the case of treatment with ICI, the phenomenon of “pseudo-progression” has been described in a variety of tumor types in which initial increase in tumor size and volume, thought to be due to immune cell infiltration, are followed by subsequent treatment response. This phenomenon suggests that there may be unique patterns of imaging changes associated with chemo-immunotherapy as opposed to chemotherapy alone.(26, 31, 32) In addition, very little is known regarding regional lymph node changes on imaging in response to neoadjuvant chemo-immunotherapy and whether increased adenopathy may be a sign of disease progression or of immune activation. In this study, we aimed to evaluate changes in regional lymphadenopathy in patients with stage 2–3 breast cancer undergoing neoadjuvant chemo-immunotherapy compared to those receiving chemotherapy alone. To our knowledge, this is the first study to evaluate lymph node changes in patients with early breast cancer undergoing neoadjuvant chemo-immunotherapy, and to explore potential relationships with surgical pathology and other patient characteristics.
We found that patients undergoing chemo-immunotherapy with pembrolizumab or pembrolizumab combined with SD-101 were more likely to develop increased lymphadenopathy compared to patients undergoing chemotherapy alone (44% vs 6.6%, p = 0.0143, Fig. 1, Table 2). Lymphadenopathy increased within the first 12 weeks of treatment prior to decreasing on subsequent therapy and occurred in a larger percentage of patients undergoing SD-101 with pembrolizumab compared to pembrolizumab alone. Of note, we found that SD-101 injections results in both ipsilateral and contralateral lymphadenopathy and may be informative for MRI monitoring of other trials utilizing localized therapies combined with systemic therapies. Similar findings have been reported in patients with non-small cell lung cancer undergoing neoadjuvant ICI who developed morphologically abnormal lymph nodes after treatment that upon biopsy were devoid of cancer cells but did show development of granulomatous inflammation.(33) Interestingly, despite early changes in lymphadenopathy, we note a persistent decrease in breast tumor size at all time points in both chemo-immunotherapy and control groups (Fig. 4, table 5). These results suggest that patients undergoing chemo-immunotherapy may be more likely to experience early lymph node changes that are independent of tumor response to treatment. One reason for this may be that chemo-immunotherapy activates immune cells residing in regional lymph nodes against tumor cells in the breast. Prior preclinical studies in mouse models have demonstrated a persistent peripheral immune activation in regional lymph nodes being associated with ongoing tumor response to chemo-immunotherapy.(34) Evaluation of patients with head and neck squamous cell cancer has previously shown an increase in CD8 + T cell activity in uninvolved regional lymph nodes in response to immune checkpoint inhibitor treatment.(35) Thus, it may be that an early increase in lymphadenopathy seen on imaging may be reflective of increased peripheral immune activation against tumor.
The overall young age of patients in this cohort (median age 45) which may lead to heightened immune activation, and thus more pronounced lymph node imaging changes on MRI, compared to older patients who typically experience immunosenesence with age.(36) Furthermore, all patients in this cohort had early stage disease and it is unclear if increasing lymphadenopathy with chemo-immunotherapy would be seen in patients with metastatic disease, particularly for those who are heavily pre-treated, as prior treatment may impair or alter the immune response to ICI. Finally, the majority of cases of increased lymphadenopathy were seen in patients undergoing intra-tumoral injections of SD-101 in addition to ICI which may have enhanced the regional lymph node response compared to pembrolizumab alone.
When looking at RCB after neoadjuvant chemo-immunotherapy, there was no significant difference between rates of RCB 0 or 1 in patients with increasing lymphadenopathy compared to those without, though the absolute percentage of patients achieving RCB 0 or 1 was higher in those with increased lymphadenopathy (66.7% vs 53.3%, p = 0.696, Fig. 2, Table 4). There were also similar rates of RCB 2 or 3 in patients with increasing lymphadenopathy compared to those without (33.3% vs 46.7%). As noted above, primary tumor parameters improved equally in both cohorts, suggesting the increased lymphadenopathy does not represent new metastatic disease/disease progression and, if reproduced in larger cohorts, does not warrant sampling mid-therapy. This is supported by the fact that, of the 12 patients who experienced increased lymphadenopathy within the first 12 weeks, 11 (92%) had subsequent decrease in lymphadenopathy over the entire 20 weeks of neoadjuvant therapy with the remaining patient experiencing stable lymphadenopathy. In addition, 11 (92%) of these 12 patients were found to have pathologically negative lymph nodes at the time of surgery. This is despite 6 of these patients (50%) having biopsy-proven positive nodes prior to starting chemo-immunotherapy. While these trends were not significant, it is possible that this is largely due to the small sample size in our study and further studies in larger cohorts are needed to explore the association of imaging lymphadenopathy and treatment response. We did note that patients undergoing chemo-immunotherapy achieved RCB 0 at a higher rate compared to control (41% vs 25%) which is consistent with larger phase III trials showing increased rates of pCR with the addition of ICI to chemotherapy.(1, 4, 37)
Our results did not indicate a correlation between estrogen receptor status, clinical lymph node status, MammaPrint score, Imprint score or tumor grade in the development of increased lymphadenopathy. We also observed similar rates of irAE in patient with increased lymphadenopathy compared to those without (66.7% in both groups), suggesting that regional lymphadenopathy may not predict development of off-target immune activity in our small sample size. This may point to differential mechanisms of immune activation associated with regional lymphadenopathy compared to development of systemic adverse events. While at this time this finding is hypothesis-generating for future investigation, it suggests the possibility of distinct immune mechanisms of response as opposed to toxicity.
As previously mentioned, one major limitation of this study is the small sample size which may lead to bias or lack of generalizability. In addition, we note that most lymph node increases were observed in patients undergoing SD-101 in combination with pembrolizumab and chemotherapy, indicating that our results may not be generalizable to larger groups of patients who undergo ICI and chemotherapy alone. Finally, MRI may not be the optimal imaging modality to assess lymph node size, as ultrasound is often the method of choice for evaluating suspicious axillary nodes.(38) However, as MRI is the best modality to evaluate neoadjuvant treatment response, findings on this modality need to be understood.
In conclusion, we found that patients undergoing neoadjuvant chemo-immunotherapy with pembrolizumab with or without SD-101 were more likely to experience early increased lymphadenopathy on serial MRI within the first 12 weeks of treatment despite concomitant decreases in breast tumor size. The findings suggest that increasing adenopathy in this clinical context is unlikely to represent disease progression and it may be reasonable to defer follow-up imaging in the neoadjuvant setting. We did not observe a correlation between residual cancer burden, clinical node status, MammaPrint high risk score, Imprint score, tumor ER status or tumor grade with increased lymphadenopathy and patients experienced similar rates of irAE regardless of lymph node changes. Future studies are warranted to determine whether on-treatment increases in lymphadenopathy may be an early sign of treatment response to neoadjuvant chemo-immunotherapy.