This case-control study compares patients who developed BC following treatment for HL with an age-matched cohort of non-irradiated patients in a 1:1 ratio. All the cases were treated at the Veneto Institute of Oncology (Padua, Italy) between 2002 and 2017.After receiving approval from the Ethical Committee (Authorization Number: CET ANV: 2023-81), we conducted a comprehensive review of data from all the patients who developed BC after receiving treatment for HL during the study period. These data were extracted from the medical records of the Veneto Institute of Oncology (IOV-IRCCS) and data collection followed identical protocols in the two groups. To establish an age-matched control group without any significant risk factors for BC, we employed a random automatic selection process. This group comprised patients who had no history of prior irradiation and lacked known genetic mutations associated with BC. The selection process was limited to patients treated at the IOV-IRCCS during the same time frame (2002-2017). To limit the bias associated with changes in treatment strategies during the years and obtain comparable follow-ups, patients were also matched according to the year of diagnosis (+/- 2 years). The same assessment methods and follow up modalities were applied to both groups. While no formal sample size calculation was performed, we included all eligible patients who met the inclusion criteria.

We compared the clinical and pathological data, treatment modalities and follow-up approaches employed in both groups, aiming to provide a comprehensive account of the management of these patients. We excluded individuals with ductal carcinoma in situ and those who were carriers of known genetic mutations that predispose them to BC.

We systematically gathered a wide range of information including familial history, physiological and pathological medical history, details regarding the age at diagnosis, neoadjuvant and adjuvant therapies administered for HL and BC, as well as pathological findings derived from surgical specimens. Additionally, we documented instances of BC recurrence, and the subsequent treatments provided. These data were recorded and continually updated in a database maintained by our research team.

These records were drawn from the clinical documentation and other pertinent sources. We applied the American Joint Committee on Cancer's (AJCC) seventh edition TNM staging system for staging. To simplify the evaluation, we categorized BC events as either "node positive" (indicating at least one positive lymph node in the definitive histological examination) or "node negative." Data for the control group were collected using the same standardized methods and were similarly entered into our database. We also compiled comprehensive information regarding HL, which encompassed details such as histological subtypes, age at diagnosis, the type of treatment employed, and the stage of the disease.

Our study defined the end of the follow-up period as the last recorded interaction with our institute, which could include the patient's last visit to surgeons, oncologists, or radiation therapists, the final radiological examinations conducted, or in cases where applicable, the patient's date of death. In most cases, the patients' follow-up assessments took place at our institute, allowing for the evaluation of clinical data through the consultation of the IOV - IRCCS clinical records. For those patients who had their follow-up assessments conducted at other healthcare facilities, telephone contact was used to obtain the necessary information. Furthermore, we measured the latency period as the time elapsed between the two diagnoses, enabling us to understand the duration between the diagnosis of HL and the subsequent diagnosis of BC. This information was crucial in assessing the potential link and the timing of these two conditions in our study.

As the risk of developing BC is greatest when patients are irradiated at young age⁶ and this may be related to a different mechanism of carcinogenesis than sporadic BC, there is some research suggesting that latency and age of HL diagnoses may impact the post-radiation therapy BCs subtypes.¹¹ Moreover, also the prior use of CT and its related menopause may alter the biological profile of BC. A dedicated analysis was performed to evaluate if the variation of these 3 risk modifying factors (namely latency, age of HL diagnoses and previous treatments) may suggest some distinctive distribution of the clinical-pathological features, the distributions of grading, proliferation index, vascular invasion, lymph node status and luminal profile.

Qualitative variables were presented as absolute values and percentages. In the case of quantitative variables, we expressed them in terms of the mean and standard deviation (SD) or median and inter-quartile ranges depending on the normality of distributions. To explore the relationships between all the relevant clinical and pathological features (namely grading, proliferation index, vascular invasion, axillary lymph nodes status, luminal profile and histological type) and the primary predictors (such as the age of HL’s diagnosis, type of treatment, and latency), we employed Chi squared tests. We utilized the Log-Rank test to compare survival curves, which helped us assess any differences in survival rates among the studied groups or cohorts. Overall survival (OS) was described according to the time between the BC diagnoses to death, while disease free survival (DFS) was defined considering the time from BC diagnoses to the first local or distant recurrence. Statistically significant P-values were considered when they were less than 0.05, and the analysis was carried out with a statistical power of 80% to ensure the reliability of our findings. We conducted our statistical analysis using STATA 11.0 software.

Hodgkin lymphoma diagnoses were made between 1963 and 2017, with 15 patients being diagnosed between 1990 and 2000, 13 before 1990, and 7 after 2000. The most common type of lymphoma was classic, with sclero-nodular type being the most common variant.

The mean age at HL diagnosis was 28.57 years (SD= 10 years).

All patients were irradiated with radiation fields involving the mammary gland. Most of the patients were irradiated using the mantle irradiation technique (93% of all patients). Radiation doses varied between 20 Gy in 10 fractions to 40 Gy in 20 fractions.

Chemotherapy was used in 25 cases (71%), with 22 patients treated with adriamycin, bleomycin, and vinblastine sulfate (ABVD) protocol, and 3 patients with the Standard V protocol. Three of them also underwent splenectomy (those patients were treated before 1990).

No significant variation in the characteristics’ distribution was observed considering latencies <20 or >20 years. Results are shown in Table 1S (supplementary material).

Moreover, the distributions of pathological findings did not show statistically significant differences according to the type of treatment for HL, but it can be observed that the ER+/HER2- and the ER-/HER2+ profiles were more frequent if CT was associated with RT (68 vs 50%, P=0.629 and 8 vs 0% P=0.383 of the cases, respectively), as illustrated in Table 2S (supplementary material).

There were no statistically significant variations in terms of distribution of pathological characteristics regarding the differences in age at HL diagnosis, except for the histotype. Lobular invasive BC was more frequent in patients irradiated for HL after the age of 30 years, compared to patients treated for HL before 30 years (5 vs 1 cases, P=0.028), as showed in Table 3S (supplementary material).

BC mean age at diagnosis was 46.3 years if HL diagnosis occurred before the age of 30 years and 52.3 years if the patients were treated for HL after the age of 30 years. Mean latency of patients treated for HL before the age of 30 years was 23.7 years, while it was equal to 14.6 years in patients treated after the age of 30 years.

In the literature, small monocentric studies have described differences in terms of immunohistochemical distributions between HL and sporadic BC populations. Dores et al.⁴ demonstrated that the risk of developing ER-/PR- neoplasia is 9 times higher for BC after HL than for BC compared to an age-matched series of BC in the general population. Horst et al.¹¹ reported that BC after HL is more often ER-/HER2- than sporadic BC (39% in 51 BC after HD, 14% in an age-matched sporadic BC group). Our data showed that ER-/HER- status was more common in the HL group (20 vs 5,7%) and ER-/PR- was more frequent in BC arising after the treatments for HL (25,7% vs 8,6%), but the differences were not statistically significant.

The results of a monocentric case series of radio-induced BC⁴,¹² indicate that the stage distribution of radiation-related BC is similar to that of sporadic BC. This is consistent with our findings and may be related to the special surveillance protocols used for these patients.

The type of treatment for HL may influence the distributions of the histological findings of BC. Indeed, irradiation may affect the number of bilateral presentations, consistent with the concept of field carcinogenesis. Our study demonstrated that this has repercussions in clinical practice, since we observed a notable difference (6 in HL vs 0 in sporadic BC group) in bilateral presentations between the 2 groups. This report finds confirmations in the literature: Yahalom et al.⁸ reported 8 bilateral BC presentations in 37 patients (22%). This should be evaluated when considering the treatment choice, and some authors¹³ have described prophylactic contralateral mastectomies as valid strategies to lower the likelihood of bilateral metachronous events.

Radiation-induced BC appears to occur in women younger than BC in the general population. A meta-analysis by Ibrahim et al.⁶ found that the mean age at diagnosis of BC after HL was 35 years, based on the data of 957 incidences of SBC in 34 studies. This suggests that the patients treated with radiation at a young age deserve a dedicated follow-up, starting before the general population. In our population, the mean age at diagnosis was 48,9 years and the global range was 32-74 years.

As described by Castiglioni et al.¹² for a small cohort of patients treated with radiation therapy for HL or other pediatric solid tumors population, the HER2 + subtype is more frequent in women irradiated during the pre-menarche period compared to a sporadic BC population. Radiation administered to the chest during breast maturation may be a risk factor for HER2 overexpression in BC, as it could trigger chromosomal instability and result in amplification of the erbB2 gene. In our group, only 6 of the irradiated patients showed amplification of HER2, but we could observe that they were diagnosed with HL at young ages (20% if the diagnosis was<30 years, 13.3 % if the diagnosis of HL was >30 years). Broeks et al.¹⁴ conducted a gene expression profiling study demonstrating that the HER+ BC in patients with prior diagnoses of BC is more frequent in patients who were irradiated at a young age. In our study the immunohistochemical surrogate profile was rare, but the mean age at irradiation was similar to the one described by Broeks et al. (equal to 22 years), suggesting a possible different mechanism of carcinogens for the patients irradiated at a young age.

The choice of the correct therapeutic approach for these patients is still controversial. At our institution, the treatment modalities used in the 2 groups were similar.

Traditionally, RT in a previously irradiated chest was avoided, so in recent decades mastectomy has been offered to these patients as the gold standard. The results of recently published studies¹⁵,¹⁶ demonstrated that reirradiation for ipsilateral BC recurrence does not increase the rate of complications. By extracting data from the Surveillance, Epidemiology, and End Results database, Burt et al.¹⁷ concluded that breast conserving therapy did not have an inferior cancer specific survival (CSS) or overall survival (OS) compared to other treatment modalities for female HL survivors who subsequently developed BC. Therefore, nowadays a conservative treatment could be acceptable even for BC arising after HL.¹⁷ The dose of prior RT (linear relationship between dose and risk of BC), age of treatment and latency may help us to select which patients may avoid mastectomy. Additionally, prior use of CT should be considered (for example, the use of anthracyclines may lead to a higher risk of BC while procarbazine is associated with a lower risk, probably because of gonadotoxicities¹⁸,¹⁹). Moreover, intraoperative RT (TARGIT-IORT) could be a reasonable option to minimize the post-RT toxicities in these populations.²⁰

The use of contralateral prophylactic mastectomy (CPM) may still have a rationale for some patients, given the elevated risk for BC and the fact that symmetry is easier to achieve when mastectomies are bilateral. The Contralateral Prophylactic Mastectomy Consensus Statement from the American Society of Breast Surgeons¹³ described patients who were irradiated at a young age as perfect candidates for CPM. In our case series, none of the 6 bilateral mastectomies had surgical complications. The results of a metanalysis⁶ show that the younger is the age at irradiation, the higher is the risk for BC. The choice of treatment modality should take into consideration: the age of treatment for HL, other treatment modalities, familiarity, and genetic predisposition.

Due to the rarity of the pathology and ethical issues, it is difficult to perform randomized clinical trials, so other retrospective and prospective studies will hopefully provide evidence for guiding the best therapeutic choices for the patients.

In our cohort, small numbers of recurrence and death events were observed, but patients with BC after HL were more likely to have local or contralateral BC recurrencies, and death events were more frequent. Regarding prognosis, the literature shows evidence from monocentric studies with conflicting data: while the overall survival (OS) is lower (also because of cardiac and pulmonary comorbidities), the BC specific survival (CSS) seems to be similar to that of sporadic BC17). A study by Moskowitz et al.⁷ demonstrated that mortality after BC was higher in childhood cancer survivors than in women with de novo BC. Some authors have proposed that undertreatment of these patients is a cause of worst OS and CSS.²¹ Therapeutic options may be limited because of prior administration of RT or CT agents such as anthracyclines.

These patients should undergo a specific and extensive follow-up, but the literature describes low awareness levels and indicates that the specific follow-up is often not offered or not performed.²² Initiation of surveillance for BC with mammography and MRI is recommended at age 25 years or ≥ 8 years from radiation for female survivors who are children, adolescents and young adults (CAYA) treated with ≥ 10 Gy chest radiation and should last up to 60 years of age.²³ Thus, awareness is a major concern for this population.

This study has several limitations, starting from its retrospective nature. The Veneto Institute of Oncology is also a reference center for radiotherapy, so most of the patients with BC arising after HL were treated for the two neoplasia in our center. However, some of the patients were treated at IOV – IRCSS only for BC, so data about specific protocols about RT for HL were not available for a limited number of patients. Another limitation is the lack of information regarding post operative complications and post-therapy toxicities (such as shoulder-arm morbidity and lung or heart adverse outcomes). The number of BCs after HL was low, partly because our study is mono-institutional, and partly because of the rarity of the pathology, despite the high relative risk for this population. HL diagnoses were made decades ago, when HL treatment modalities were very different from current practice. Historically, RT was extensively used in the HL treatment protocols, but recent strategies have minimized its use, favoring CT over RT and limiting RT indications, with RT often omitted in favorable HL. However, we believe that, considering the decades of latency, the results of this study are still of current interest, because these are the cohorts of patients who are currently being treated for BC. Moreover, the rate of secondary BC is still elevated, even in cohorts of HL patients with lower radiation exposure.⁵