Interventional oncological treatment of breast cancer liver metastasis (BCLM): single center long-term evaluation over 26 years using thermoablation techniques like LITT, MWA and TACE in a multimodal application

Abstract The purpose of the study is to retrospectively evaluate the development and technological progress in local oncological treatments of patients with breast cancer liver metastasis (BCLM) using LITT (laser interstitial thermotherapy), MWA (microwave ablation) and TACE (transarterial chemoembolization) ablation techniques in a multimodal application. The study uses data generated between 1993 and 2020. Therapy results were evaluated using the Kaplan–Meier survival estimate, Cox proportional hazard regression and log-rank test. Cox regression analysis showed that the different treatment methods are statistically significant predictors of survival of patients. Median survival times for groups treated with LITT (212 patients) and LITT + TACE (215 patients) were 2.2 years and 2.1 years respectively; median survival times for groups treated with MWA (17 patients) and MWA + TACE (143 patients) were 5.6 and 2.4 years respectively. For LITT only treatments, the 1-, 3- and 5-year survival probability scored 80%, 37%, 22%. Results for combined LITT + TACE treatments were 76%, 34% and 15%. In group MWA, the 1-/3-/5-year survival probability rates were calculated as 89%, 89%, 89% (however, they should be interpreted carefully due to a relatively small sample size of n = 17 patients). Group MWA + TACE offered values of 77%, 38% and 22%. A separate group of 549 patients was analyzed with TACE monotherapy treatment. The estimated median survival time in this group was 0.8 years. The 1-/3-/5-year survival probability rates were 37%, 8% and 4%. Treatments with combined MWA and MWA + TACE resulted in the best median survival time estimations in this study.


Introduction
Breast cancer (BC) is the most commonly diagnosed cancer in women [1][2][3].BC is described as a heterogeneous disease with specific molecular subtypes, which are associated with different prognoses and responses to different treatments.Approximately 50% of all patients diagnosed with BC develop metastases.Common metastatic sites are lung, liver, brain and bones [4,5].Liver metastases resulting from breast cancer (BCLM) develop in approximately 50% of all patients with metastatic BC and rank among the most common first sites of spread [5] In advanced stages of breast cancer, metastases restricted to the liver develop in approximately 5-12% of the patient [6].BCLM has been identified as an indicator of the presence of disseminated cancer with a very poor prognosis even if it appears to be limited to a single organ.
BCLM rarely presents itself with symptoms until the liver function deteriorates beyond a certain threshold [7].Surgical approaches, like liver transplantation or resection, are pursued.However, only a minority of cases fit the necessary requirements.Furthermore, surgical resections of liver metastases from breast cancer are still a subject of discussion.
Due to the promising results of percutaneous and laparoscopic ablative technologies in colorectal liver metastases (CRLM), studies reported on the application of these methods for the successful treatment of BCLM [8].Compared to surgical resection, minimally invasive curative treatments are cost-effective, tissue conserving and provide a relatively low complication rate [5] Successfully applied percutaneous thermal ablation treatment methods for unresectable hepatic malignancies are radiofrequency ablation (RFA), laser interstitial thermal therapy (LITT) and microwave ablation (MWA).
Among these loco-regional treatment modalities for unresectable hepatic disease, radio-frequency (RFA) technology advancements became stagnant.RFA is not considered in this study due to the size of patient groups which suited our inclusion requirements.However, it has been shown that RFA is a successful treatment method for certain BCLM therapies [9].MR-guided LITT has been a common treatment method in the last two decades.It can be considered a safe and effective treatment for well-selected patients [10].LITT uses an Nd:YAG laser emitting light with a wavelength of 1,064 nm with cooled applicator fibers in continuous wave mode.Multiple applicators and pullbacks can be used to enlarge the ablation zone.The main advantage of LITT over other ablative methods is that the ablation procedure can be accurately monitored using thermosensitive T1 weighted MRI sequences, while the main disadvantage of LITT is that it is a time-consuming procedure since the sheath for the laser fiber should be introduced under CTguidance, then the patient needs be transferred to the MRI for the ablation procedure.The main indication for LITT is patients with five or fewer liver metastases with a maximum size of 5 cm each.
Microwave ablation (MWA) is based on an electromagnetic field (0.9-2.450 GHz), that radiates from a tip of an antenna.Due to a high content of water molecules in the tissue, microwaves induce atomic and dipole rotation resulting in friction among them.The agitation of water molecules increases the temperature, which finally induces tumor cell death via coagulative necrosis [11] Thermal ablations are limited to a certain volume of the necrosis ablation zone.In contrast to RFA, microwave probes provide faster tissue heating over a larger volume with a less prominent 'heat sink effect' [12].
In the treatment of BCLM, several studies have also confirmed the efficacy of transarterial loco-regional treatments including TACE (transarterial chemoembolization) which may be applied in selected patients with chemo-resistant advanced metastatic liver disease [8].TACE has been identified as an effective therapy for neoadjuvant, symptomatic, or palliative therapy indications.In TACE, high doses of chemotherapeutic agents are directly delivered to the target tumor [13,14].TACE may be applied as monotherapy; it may also be combined with thermal ablation treatments, or such treatments can be applied exclusively.TACE is performed in several treatment sessions in order to achieve complete devascularization of all tumors while protecting non-affected parts of the liver.
Regarding treatments of BCLM at our University Hospital, which are considered in this study, LITT was mainly performed between the years 1993 and 2010.The number of MWA treatments has increased since the year 2010.TACE treatments were first applied in 1999.During the course of more than 26 years of treatments between 1993 and 2020, MWA has then replaced LITT in the treatment of BCLM.The purpose of the study is to retrospectively evaluate the development and technological progress in local oncological treatments of patients with BCLM by means of different interventional therapies within a long observation period and their effect on patient's survival.The main aim of the study is to report the outcome of the different interventional treatment methods in terms of patients' survival.

Materials and methods
The current retrospective study was approved by the ethical committee of the Frankfurt University Hospital, and registered under the number 2021-2022.All patients included in the current study signed informed consent before each treatment.The patients were also informed about the possible anonymous usage of their data and imaging studies for research purposes, which they approved.The study included treatments of patients at our hospital between the years 1993 and 2020.

Patients' selection criteria
As an initial step, the database had to be filtered regarding consistency and completeness of key selection criteria.The main data elements were the availability of the dates of the first and last treatment, and the date of last contact or death.Data regarding tumour stage or other information about cancer-relevant health status of patients were not available or were not consistently reported in the database and therefore could not be considered in this study.
The analysis of our dataset concentrates on patients with breast cancer as the only known primary tumor and breast cancer liver metastases as the only ablation treatment region.Patients with either extrahepatic metastases or other known primary tumors were excluded.Narrowing this field of analysis allows focus; however, it limits the number of cases in our database.Within this entity of datasets, five groups were selected in order to analyze different types of relevant treatments or combinations of treatments.We differentiated patients who were treated with LITT, LITT þ TACE, MWA, MWA þ TACE or TACE as monotherapy.All five groups of patients were included in the current study including patients who received only TACE.
A very small number of patients received both LITT and MWA during the transfer phase of the ablation from LITT to MWA in our department.This small number of patients was excluded to avoid overlap and inconsistency in the analysis.

Treatment selection criteria
Regarding this long observation period, it has to be considered that treatment methods did evolve over time i.e. chronologically.LITT was applied in earlier phases followed by MWA in later phases.Hence, the decision to treat using LITT or MWA ablation was not based on specific selection criteria related to the patient stage or size of the tumor, but the decision to ablate regardless of the method used was based on the size and number of metastases, namely 5 or fewer metastases and none of them larger than 5 cm.For the metastases that did not meet the ablation criteria, TACE was used as a neoadjuvant/downstaging procedure to reduce the size, and less likely, the number of metastases to achieve the indication criteria for ablation.It was also used to reduce tumor vascularity and maximize the ablation effect in large lesions.Cases with failed adequate down staging under TACE, remained as TACE monotherapy.TACE as monotherapy has been selected for comparison reasons.This group needs to be interpreted carefully since TACE treatments as monotherapy were clearly performed in patients with resistant and advanced stages of liver metastases as in patients with TACE and ablation or ablation only.
As the first long-time analysis based on our dataset, this study, therefore, excludes further multivariate analysis of patients' status.However, it does consider the impact of age, and statistically analyses the initial key question of whether different treatment methods affect survival.The following steps may evaluate possible model extensions including other data sources and further in-depth analysis.

Statistical methods
Raw data has been extracted using Microsoft Excel.Excel tables were used as input files for statistical analysis using SPSS, Version 22 as well as R package version 4.1, especially software packages ,,survival "and ,,ggplot2" [15,16].Overall survival rates (OS), as our primary outcome, were calculated using the Kaplan-Meier survival estimate [17].
The starting point for the calculation of the estimator was the date of the first treatment session at the institute.The end date for the calculation was the date of death or the last contact with the patient, which could either be the last ablation session at the institute or a checkup date.Differences between groups of patients based on treatment modalities were analyzed using the Cox-Mantel log-rank test and the Kaplan-Meier survival estimate.A p value 0.05 was considered statistically significant.Cox regression was used as a prediction method in order to analyze relations between several predictors of survival [18].
Risks of death were marked by means of hazard ratios including 95%-confidence levels.

Results
A total number of 212 patients were treated with LITT (377 treatments, all registered as female).The group consisted of patients with ages between 28 and 81 years (median age 57) (Table 1).Regarding the combination of LITT þ TACE, ablation with LITT was applied 430 times on 215 patients.Median age in this cohort was 57 (from the age of 29 years to the max.age of 84 years).Within the LITT þ TACE group, an average of 5 to 9 TACE-cycles were performed (1,266 cycles total).
The total complication rate per treatment for LITT group was 17.0%; LITT þ TACE led to results of 12.3%.In groups LITT and LITT þ TACE, most complications were pleural effusion and subcapsular hematoma.In both MWA groups, relatively few complications were registered.
The first data in this study were LITT treatments between the years 1993 and 2010.For this patient selection, the first MWA treatment was in 2010 and the last treatment in this cohort during the chosen time span was recorded in April 2020.TACE has been performed between the years 1999 and 2020.
Regarding MWA, our study includes only 17 patients who received 24 MWA treatments: A relatively small sample size compared to the other four groups in this study.One hundred forty three patients received MWA combined with TACE.In this group, patients with TACE were treated with an average of 2.6 microwave ablations per patient in combination with 7.1 TACE cycles per patient (1,016 cycles total).The MWA group showed a value of 1.4 treatments per patient.Patients in group MWA and MWA þ TACE were of similar age: median age 51 vs. 55 years and min./max.age: 28/77 vs. 28/78 years.
In our study, we analyzed 549 patients solely treated with TACE.This method reached the highest amount of 2239 treatments in total.This group scored a median survival time of 0.8 years.On average, patients were treated with TACE 4.1 times.
In the following Figure 1 Kaplan-Meier graphs are depicted for all 5 groups: MWA, LITT, MWA þ TACE, LITT þ TACE, and, TACE monotherapy.Table 2 provides the number of patients for reference in the first row.The number of events indicates cumulative incidence rates, i.e. cases of death.Survival times (years) are shown as mean and median values (median with upper and lower confidence intervals).The following sections of Table 2 contain 1-/3-and 5-year survival rates (%) with 95%-confidence intervals by treatment method, calculated based on Cox regression.Figure 1 summarizes the timelines for the cumulative incidence rates, i.e. events, by treatment method."Dots" along the curves represent censored cases, i.e. patients that have left the observation/the study.Several dots may overlay at one point in time.The table attached below lists the exact number of included patients "at risk" along the timeline, the occurring events (death of patients) as well as the number of patients when the observation ended (either date of death or last date of contact).
Complications for MWA occurred in 4.2% of all treatments in this cohort.For MWA þ TACE, the complications rate was 1.9%.Complications in both groups were pleural effusion, subcapsular hematoma, abscess, intrahepatic bleeding, and intestinal fistula.
An example of a combined TACE and MWA treatment is shown in Figure 2. The MRI scans in Figures 2(a and b) highlight the liver metastasis before partial remission.Figure 2(c) depicts the CT-guided biopsy and Figure 2(d) shows the TACE treatment as an angiography of the liver.The effects of chemoembolization can be seen in Figures 2(e and f) during the regular MRI check, after 4 months from TACE treatment.Figure 2(g) contains the CT scan of the MWA treatment, which has been performed 7 months after TACE treatment.A reduction of the metastases as a result of the combined TACE and MWA treatment is presented in Figure 2(h Therapy results were evaluated using the survival rates of the Kaplan-Meier estimator.Table 3 provides a descriptive comparison between pairs of Kaplan-Meier curves for all possible therapy combinations using a log rank-test.P-values 0.05 indicate statistically significant differences in survival rates between the treatment methods within the descriptive comparison of pairs.As a result, rows a, d, g, i, j show p 0.05, and consequently a statistically significant survival rate difference according to this level of significance.
For example, LITT vs. MWA with p ¼ 0.166, LITT vs. MWA þ TACE with the highest p-value of even 0.817 or LITT þ TACE vs. MWA þ TACE with p ¼ 0.370 and LITT þ TACE vs. MWA with 0.063 are above p-level of 0.05 and therefore show no statistically significant difference.Thus, MWA cannot fully show its "potential" regarding this test in this cohort.
Regarding the TACE-monotherapy comparison with other treatment methods in this list, significant differences are obvious: For example, comparing MWA with TACE or MWA þ TACE vs. TACE leads to a p-value < 0.001.As a result of this specific analysis, it can be stated that the chosen treatment method remains a statistically significant predictor for survival rates (considering p 0.05), even when age of patients and sex is added to the model.In other words, treatment methods adjusted to age and sex are relevant for the survival of patients.It proves that the survival/death rates of this study are not mainly driven by increasing age.Sex plays no or a very minor role in this cohort since only 4 out of the 1136 registered patients are males.
Table 5 shows the predictive values of the different treatment methods, not adjusted by age of patients and sex (f/m) as multiple univariate models (i.e. one statistical model per method).Each chosen method is compared with the rest of all other methods (without the chosen method) and tested regarding the prediction of survival (e. g.LITT only vs. non-LITT only etc.) The age of patients and sex as potential predictors are not included in this statistical test.
A p-value 0.05 indicates statistically significant survival predictors (see shaded rows: MWA, MWA þ TACE and TACE).This is the case for all five methods included in this study (shaded rows in the table).
Taking the example of LITT þ TACE, the hazard ratio of HR ¼ 0.787 indicates a 0.787-times reduced risk of nonsurvival, i.e. risk of death, in comparison with all other treatment methods applied to the total number of patients (95%-CI between 0.668 and 0.928).
The following table (Table 6) includes the Cox regression test, adjusted by age of patients and sex (f/m) as a multivariate statistical model.Adjustments allow the calculation of statistical effects of additional potential survival factors such as age and/or sex.
As potential predictors for survival, and due to restrictions of the algorithm, 4 ablation methods (LITT, LITT þ TACE, MWA and MWA þ TACE), as well as the age of patients and sex (female), are included "at the same time" in this multivariable model; TACE had to be excluded.The test shows the statistical significance of the chosen treatment method while at the same time considering the age and sex of patients.As a result of this test, it can be stated for the selected population of this study that the 4 treatment methods included are relevant for the prediction of survival.Statistically, age is not a significant predictor for the death of patients, based on the set p-value of 0.05.As stated, sex has no relevance in this cohort.
Hazard ratios again relate to the risk of death, i.e. the non-survival of patients in this study.For example, the hazard ratio for LITT (first column of Table 6) with a value of HR ¼ 0.314 indicates a death risk factor of 68.6% (1-0.314¼ 0.686, 95%-CI value between 0.26 and 0.381).This  means that the risk of death in the case of treatment with LITT is 68.6% lower compared to all other factors included in the model.Another example MWA shows a hazard ratio of HR ¼ 0.16.Thus, it can be stated that MWA has a risk-ofdeath which is 84% better compared to all other factors (1-0.16¼ 0.84).Based on this statistical test, amongst the compared methods MWA ranks best (HR ¼ 0.16), followed by LITT (HR ¼ 0.314) closely followed by MWA þ TACE (HR ¼ 0.319), LITT þ TACE (HR ¼ 0.387).Regarding the age of patients, HR ¼ 1.006 indicates that for each þ1year-step along the timeline, the risk-of-death increases by the factor of 0.6 (95% CI between HR-levels of 0.999 and 1.013)

Discussion
To our knowledge, this is the first retrospective single-center study over a relatively long time period of more than 26 years.The analysis is based on a range of patients and treatments, which were performed between 1993 and 2020 at our University Hospital.Within this observation period, treatment concepts toward BCLM have evolved.This paper aims to point out the achievements made in interdisciplinary therapy management over the past 26 years, considering the survival rates of patients with BCLM.
Treatments with MWA-only resulted in the best median and mean survival time estimations, followed by MWA þ TACE in combination.The 3-and 5-year survival rates of MWA and MWA þ TACE treatments are higher compared to both LITT groups, demonstrating that these methods appear to be promising modalities for the treatment of unresectable BCLM.Here it is important to take into consideration the previously mentioned treatment protocol.Hence, patients who received LITT only or MWA only were patients who had 5 or fewer lesions none of which is larger than 5 cm at first presentation.This group of patients clearly represents a different stage from those who received LITT and TACE or MWA and TACE and who had lesions larger than 5 cm, or had more than 5 lesions or both, and the lesions were downsized using TACE.The TACE-only group represents the most advanced stage of patients since those patients were either resistant to downsizing or had a large number and/or large size of metastases from the beginning.However, statistical comparisons of MWA versus other treatment methods have to be interpreted carefully due to the relatively low sample size in this group.
In comparison to the group of patients who were treated with TACE as monotherapy, patients who were treated with MWA, LITT or in combination with TACE showed significantly higher survival times.As stated, this group should be considered with special attention, due to patients' characteristics.It has been included to offer results within the array of relevant applied methods and the statistical significance of each method on survival.
The success of treatments by means of resection or ablation depends on multiple factors.Key factors are the size and number of tumors as well as the degree of liver dysfunction.It should be considered that treatment methods regarding apparatus' (e.g.probes, fibers etc.), as well as imaging technology, have developed over time, and therefore may affect results.Comparisons have to be interpreted carefully also due to the development of chemoembolization.Indepth analysis in the next steps could therefore focus on each modality and the impacts on efficacies, i.e. survival rates.Further evaluation could include patients' comorbidities and overall status, especially the function of the liver, as well as individual characteristics of the BCLM.Differentiation by treatment method, more specific conclusions regarding efficacy, i.e. survival rates per sub-groups could be developed by means of further multivariate analysis, where data is available.However, the limitation of statistical methods regarding the sizes of cohorts and sub-groups needs to be considered in advance.
One of the main limitations of the current study is the lack of data regarding the size, number and location of the treated lesions in the liver.The main reason is the lack of imaging studies for patients performed before full digitalization of our department, those studies were either given to the patients as hard copies, stored as hard copies in archives or destroyed after exceeding the legal obligatory storage duration.Still, due to the consistent protocol of ablation in our department over the years regardless of the ablation method used (maximal diameter 5 cm and maximum number 5 lesions, lesions outside these criteria received TACE for downsizing to meet the ablation criteria and if downsizing failed the patients remained on TACE only), it can be assumed that patients who were treated with LITT had similar tumor criteria as those treated with MWA since both methods were used chronologically and not at the same time (except for a very short overlap time).The second limitation is that we included a group of patients with more advanced and/or resistant lesions namely the TACE-only group since this represented patients who were outside the ablation criteria or with failed downsizing.However, this group was included to complete the spectrum of our study, which aimed at addressing the long-term survival (26 years) following interventional management of breast cancer liver metastases.Further studies comparing these interventional procedures in randomized studies, while taking into consideration detailed tumor criteria, are required to overcome limitations of the current study.A further limitation of the study is its retrospective nature and the fact that different technologies over a relatively long period of time were included in the analysis.However we still think that summarizing all interventional treatment modalities offered to a specific disease entity in a single study gives an insight into the technological developments and improvements and their impact on patients' care.This study has statistically proven that survival rates of patients treated with different ablation methods do not simply depend on increasing patients' age.We could show that the chosen treatment method has a statistically significant influence on the survival of patients.Patients who received MWA only had better survival than those who received LITT only followed by MWA þ TACE and LITT þ TACE.The least survival was observed among the TACE-only group of patients.This conclusion should however take into consideration the previously mentioned selection criteria for treatment.Further clinical relevance of MWA may be expected in BCLM management owing to its multiple advantages.However, prospective data remain necessary to further evaluate the superiority of either modality.

Figure 1 .
Figure 1.Kaplan-Meier estimator for different treatment methods MWA, LITT, MWA þ TACE, LITT þ TACE, TACE and list of events along the timeline.Data for the number of patients at risk, number of events (death of the patient) and number of patients where the observation ended (obs.end).
) by means of the MRI scan, taken 1 month after MWA treatment.Initial MRI of the liver with T2 weighted image with fat saturation (A) and T1 weighted image (B) showing multiple liver metastases before treatment.CT image (C) showing CTguided biopsy of one of the lesions.DSA (D) during downsizing of the lesions using TACE.T2 (E) and T1 (F) weighted MRI following 4 sessions of TACE showing downsizing of the lesions.CT guided ablation (G) of the residual lesion in Segment 7 of the liver after 7 months from the beginning of treatment.Follow-up MRI (H) one month after MWA ablation showing the ablation zone in the liver Comparing 3-and 5-year survival rates of MWA-based treatments with other treatment methods improvements are obvious.However, this should be interpreted carefully due to the limited sample size of MWA-only treatments.Less postoperative complications underline the success of MWA.

Figure 2 .
Figure 2. Treatment of the patient with BCLM with combined TACE and MWA.

Table 1 .
Characteristics of patients and results.
Table 4 includes results of the Cox regression test as a general test including all treatment methods adjusted by age of patients and sex (share of female patients).A p-value 0.05 indicates a statistically significant survival predictor.

Table 3 .
Descriptive comparison between pairs of Kaplan-Meier curves using log rank-test.

Table 5 .
Cox regression test not adjusted by age of patients and sex (f/m) as multiple univariate model.