Seroma and Radiotherapy.

SEROMA AND RADIOTHERAPY

Carl Stuart

Stuart Medical Series

Paper Format: AIP Style.

Table of Contents.

1.     Abstract

2.     Introduction

3.     Background

4.     Literature Review

5.     Methodology

6.     Results

7.     References

Abstract

Breast cancer is the second most prevalent cancer type in women. Current treatment modalities for it include chemotherapy, radiotherapy, surgical ablation and combination therapy. One of the most effective forms of combination therapy is breast conserving which consists of subtotal breast-conserving surgery and post-operative irradiation. Various post-operative radiotherapy techniques are available. In this study, whole breast radiation therapy is used, and the main aim is to determine the extent of seroma reduction during whole breast radiation therapy and to verify the improved dose distribution of the resulting plan when a patient is re-simulated - using the XIO treatment planning system - due to the presence of a large postoperative seroma.

Introduction

Breast cancer is the second most prevalent cancer in women, with the most prevalent one being skin cancer. Statistical studies have shown that 30% of all cancers diagnosed in women are breast cancer. Its incidence has continuously increased since 1975. Statistical analyses also forecasts that 12% of the total US female population will be affected by breast cancer⁶.

Ductal carcinoma in situ (DCIS) makes up 85% of non-invasive breast tumors. Epidemiological studies have shown that breast cancer causes approximately 40,000 deaths per annum. Therefore, the mortality rate is classified as being second after lung cancer⁶.

The risk factors that predispose women to breast cancer are discussed hereafter. First of all, a positive family history of the diagnose increases the probability. Secondly, women whose genotype includes BRCA1 and BRCA2 genes have approximately 80% probability of developing breast cancer. Moreover, there is a positive correlation between these genes and increased incidence of ovarian cancer. However, for women lacking the BRCA1 and BRCA2 genes, spontaneous genetic mutations do predispose them to the disease. Another risk factor of developing breast cancer is age; this is as studies have shown that breast cancer has a predilection for post-menopausal women⁶.

Since the 1990s, awareness campaigns and effective interventions (that have been used to manage breast cancers) have resulted in a reduction in mortality rate by 2.5% in white women and 1.4% in black women. Moreover, in 2011, the number of survivors of breast cancer was estimated to be over 2 million. The awareness campaigns encouraged more women to be screened for breast tumors. The commonly used screening methods are physical breast examination and mammography. Such screening enabled the health care providers to diagnose the early-stages of the cancer. Such forms of breast cancer are amenable to the therapies that effect complete cure, thereby reducing the mortality rate associated with breast cancer⁶.

Current treatment modalities for breast cancer are chemotherapy, radiotherapy, surgical ablation and combination therapy. One of the most effective forms of combination therapy is breast conserving therapy which consists of subtotal breast-conserving surgery and post-operative irradiation. The surgery removes the bulk of the tumor and the irradiation causes the death of the remaining residual neoplastic cells, thereby reducing the probability of cancer recurrence⁶.

Radiation therapy is a recognized modality used for treating the disease. Initially, whole breast irradiation was performed using 2D (two dimensional) techniques with dose distribution calculations being done using the central-axis axial plane. However, these calculations provided inadequate information with regards to off-axis planes. Additionally, tissue inhomogeneity correction was never accounted to using 2D techniques, and this resulted in some segments of breast tissue receiving radiation doses in excess of the prescription dose. Fortunately, the advent of CT (computed tomography) scanners enabled 3D (three-dimensional) images to be utilized during the treatment procedure, thus accounting for inhomogeneity correction in soft tissues. Moreover, dose calculations utilized complex algorithms that accounted to all axial planes. Therefore, CT-enhanced radiotherapy was precise and more effective than the conventional radiotherapy³.

External beam radiation cannot achieve dose homogeneity due to differential tissue densities and uneven body contour. This has necessitated the development of beam modifiers that are capable of distributing radiation dose uniformly. During breast radiotherapy, tangential field irradiation results in dose inhomogeneities because the breast tissue is thin anteriorly and thick posteriorly. Moreover, its conical shape alters the radiation dose distribution. To compensate this drawback, the wedge technique is used in breast irradiation. However, even wedge techniques do not achieve conformal dose distribution, and they still create coldspots and hotspots within the tissue. Thus, the wedge technique does not eliminate the toxicity associated with breast irradiation, and this has led to the development of more effective radiation techniques such as the FiF (field-in-field) technique and the iIMRT (inversed-planned intensity modulated radiation therapy). FiF technique utilizes sMLC (static multi-leaf collimator) subfields to block hotspot areas thus improving dose distribution. Its multi-leaf collimator is constantly static while the beam moves accordingly as planned. The iIMRT technique utilizes DVH (dose volume histogram) to determine inverse planning. The technique also optimizes the constraints by contouring critical structures and determining the PTV (planning target volume) ³.

The purpose of this study is to determine the extent of seroma reduction during whole breast radiation therapy and to verify the improved dose distribution of the resulting plan when a patient is re-simulated - using the XIO treatment planning system - due to the presence of a large postoperative seroma.

Background

Post-operative radiotherapy for breast cancer does improve the prognosis of the disease by augmenting local control of the neoplastic cells. It has been proven that the radiation boost that is administered following the whole breast irradiation reduces the incidence of tumor recurrence. Studies have also shown that the extent of seroma reduction in surgical ablation patient ranges between 36-50%, while for patients who have undergone both surgical ablation and post-operative irradiation, the extent of seroma reduction is about 62%. As a result of seroma reduction, the volumetric proportion of normal breast tissue increases in relation to the whole breast tissue volume; and if the normal tissue is irradiated, there is an increased risk of it undergoing fibrosis, thus altering the cosmetic outcome. It, therefore, follows that good cosmetic outcome results appear when the irradiated normal volume is as small as possible³.

Non conclusive studies have been done on the dosimetric outcome of the various planning techniques for post-operative seroma irradiation. Currently, the planning of boost irradiation can be done separately or integrated with breast irradiation, and such planning confers the advantage of allowing a CT scan to be obtained prior to the planning of irradiation³.

Large seroma reduction is mandatory in the cases of large initial seroma volumes, and this usually results in minimizing irradiation overdose during the sequential boost radiotherapy¹. Studies have also shown that simultaneous integrated boost confers the dual advantages of improved dose distribution and short treatment course. The dosimetrician or the dosimetrist must, therefore, know which technique and plan results in the minimal excess irradiation dose outside the PTV boost in order for him or her to determine the most optimal technique that can be used for a particular seroma patient³.

Literature Review

In 2011, Alderliesten et al published a study entitled Dosimetric impact of post-operative seroma reduction during radiotherapy after breast-conserving surgery in which the effect of seroma reduction on radiotherapy was studied. The aim of the study was to compare three boost RT techniques and evaluate their dosimetric effect on seroma reduction during post-operative irradiation following breast-conserving surgery. The subjects of this study were 21 patients who had developed seroma after breast-conserving surgery. Three CT scans were done for each patient: one prior to the surgery and the other two during post-operative radiotherapy. The following three radiation plans were also generated for each patient: in the first plan, the whole breast irradiation is done one week prior to radiotherapy, with the sequential boost being given at the fifth week during radiation therapy; in the second plan, simultaneous integrated boost (SIB) was given one week prior to radiotherapy, and, in the third plan, SIB-adaptive radiation therapy that was planned to be done one week prior to radiotherapy and also re-planned to be done three weeks through the radiotherapy. During the fifth week of radiotherapy, the irradiated volumes, maximum heart dose and the mean lung dose were calculated and compared. The results showed that RT caused 62% seroma reduction with SIB-adaptive radiation therapy being more effective than the other two plans in reducing the volume of breast tissue receiving excess radiation. Thus, it can be concluded that this study has shown that simultaneous integrated boost radiotherapy has confered dosimetric advantages to seroma patients on RT¹.

In 2009, Sharma et al published the findings of a clinical investigation they had conducted under the title Change in seroma volume during whole-breast radiation therapy. The investigation evaluated the dynamic changes in breast seroma volume during whole-breast radiation therapy. In this investigation, retrospective reviews were done on women receiving BCS-therapy for seroma. All the subjects received either a standard fractionated regimen or a hypofractionated regimen. During the investigation, CT simulation was done prior to the surgery and then prior to the boost irradiation. This enabled the investigators to contour and compare the seroma volumes. The results of the investigation showed that the mean seroma volume prior to the whole-breast radiotherapy, and at the time of boost planning it showed a significance difference. Calculations showed that there was about 50% reduction in volume from the initial seroma level. Moreover, the investigation revealed that the fractionation schedule showed no correlation to the seroma volume changes. However, there was an inverse correlation between the changes in seroma volume and the duration from surgery to commencement of radiotherapy. Therefore, it can be concluded from this investigation that whole-breast irradiation causes significant changes in seroma volume, and this, in turn, affects the accuracy of the boost planning. Also, it can be concluded that CT-based boost planning has ensured appropriate dose coverage⁷.

In 2010, Yang et al published their study under the title Planning the breast boost: How accurately do surgical clips represent the CT seroma?. This study used clip-based electron fields to evaluate the dosimetric coverage of various CT-defined radiation boost volumes, and to perform distance measurements between the edge of the CT-defined seroma and the surgical clips in the coronal plane in patients who had undergone wide local breast cancer excision. The study reviewed the planning CT scan images of thirty lumpectomy cavities (of 30 female patients), with each seroma cavity having a minimum of 4 clips and a cavity visualization score of P3. Distance measurements were done as per the set specifications of the study, and the CT tumor beds were used to devise the 3D conformal radiotherapy plans. The analysis of the dose-volume histogram (DVH) parameters of the boost treatment plans was also done. The results obtained in this study showed that the mean seroma edge was about 0.5 cm beyond the clips. About 46.7% of the patients had a D90 and a geographical miss of 36.7% when the electron fields were used. Thus, it can be concluded from this study that surgical clips showed no consistency with the seroma edge and that the electron boost field had led to inadequate dose coverage in CT tumor beds. However, based on the findings, it can be deduced that 3D conformal radiotherapy plans should be utilized during the determination of the dose coverage for the tumor bed⁹.

In 2013, ZhaoZhi et al also published another study under the title Simultaneous Integrated Boost in Breast Conserving Radiotherapy: Is Replanning Necessary Following Tumor Bed Change? This study evaluated the changes in seroma volume using repeat CT scans and also exploring for the necessity of re-planning in BCR using IMRT with SIB. Thirty patients were involved in the study, and they all underwent whole breast irradiation and were, thereafter, subjected to boost IMRT-SIB scans at the sixth week of radiotherapy treatment. The appropriate volumes were delineated in the CTs and compared. Both re-plans and hybrid plans were also constructed and utilized in the study. The results showed that there was a reduction of about 40% in the mean tumor bed volume. There was also adequate target coverage in both hybrid plans and re-plans. The results also revealed that re-planning reduced the mean dose of the whole breast. It can thus be concluded from this study that a significant reduction of the tumor bed volume can be achieved by a fractionated schedule of IMRT-SIB with the boost irradiation managing the remaining tumor volume. It is also apparent that re-planning avoids the unnecessary use of high dose irradiation outside the boost regimen¹⁰.

In 2008, Kader et al conducted a clinical investigation which was later on published under the title When is CT-based postoperative seroma most useful to plan partial breast radiotherapy? Evaluation of clinical factors affecting seroma volume and clarity. The investigation assesses the effects that clinical factors and the time interval from surgery had on seroma volume and to establish the most apposite time to utilize the CT-based seroma to plan PBI (partial breast irradiation). This clinical investigation studied 205 female patients who were afflicted by early-stage breast cancer and had undergone planning CT after BCS. The seroma volume was contoured and the seroma clarity scored using the conventional Seroma Clarity Score. Both multivariate and univariate analyses were used to evaluate the associations between seroma characteristics and the time interval from surgery and related clinical factors. The results showed that the mean time interval from surgery to the first CT was 84 days. Also, during the postoperative weeks, both mean seroma volume and seroma clarity score gradually decreased. Moreover, both showed a significant correlation to the excised breast tissue volume, but they were not correlated to the initial tumor size, chemotherapy use or surgical re-excision. It could thus be concluded that this clinical investigation showed that 8 weeks post-operation was the optimal time for obtaining the planning CT scans for PBI. Also, it can be deduced that alternate strategies for identifying PBI targets should be sought after the fourteenth post-operative week. Additionally, it can be inferred that CT-based seroma must never be used as the only guide for PBI volume definition⁴.

In 2012, Kosztyla et al published a study entitled Evaluation of Dosimetric Consequences of  Seroma Contour Variability in Accelerated Partial Breast Irradiation Using a Constructed Representative Seroma Contour. The study aimed to quantify the extent of the dosimetric impact of observer contouring variability on APBI (accelerated partial breast irradiation) through the construction of a RSC (representative seroma contour). The study involved 21 patients who had 4 CT-scans of their APBI-amenable seroma taken and it was contoured after each CT scan had been taken. Repeat contouring was done to assess intra-observer variations, and the constructed RCS was used to quantify the seroma contour variability. PVO (percent volume overlap) and the RMS (Root-mean-square) differences were also calculated. Treatment fields were re-applied to repeat CTs. The results showed that inter-observer RMS differences were greater than the intra-observer RMS differences. The parameters used to assess for dosimetric impact were the same for each and every patient, and they showed that their median RMS differences were minimal. Thus, it could be concluded from this study that inter-observer variations were the significant cause of seroma contour variations. Moreover, it is apparent from the study that the planning margins used provide adequate dose coverage for the seroma despite the contour variations⁵.

In 2011, Chadha et al published a study entitled Image guidance using 3D-ultrasound (3D-US) for daily positioning of lumpectomy cavity for boost irradiation. The study evaluated the utility of 3D ultrasound breast IGRT (Image-guided radiotherapy) for photon and electron lumpectomy-site boost radiotherapy. The study enrolled 20 patients who were on either electron or photon boost. 3D ultrasound images that were acquired were used to construct a co-registered CT/3DUS (3D ultrasound) dataset. Intrafractional motion parameters were calculated and the IGRT shift determined. The photon shifts were evaluated isocentrically; while the electron shifts were examined using the beam eye-view. Volume differences between the boost and simulation fractions were calculated. Other IGRT shift parameters were also calculated. The results showed that the IGRT shifts for photon boosts ranged between 0.6 – 1.5 cm with 50% of the fractions requiring a shift of more than 1 cm. There was also a significant volume change between the boost and simulation volumes. IGRT shifts for electron boosts ranged between 0.6 – 1.5 cm with 52% of them extending out of the dosimetric penumbra. Interfraction analysis also showed that a significant portion of the shifts extended out of the dosimetric penumbra. It can thus be concluded that significant lumpectomy cavity shift occurs during fractionated radiotherapy and that 3D ultrasound imaging can be used to correct the interfractional motion. However, this study has demonstrated that further studies are needed to clearly define the appropriate protocols for clinical use of IGRT in the management of breast cancer².

In 2013, Yue et al published a study entitled Tracking the dynamic seroma cavity using fiducial markers in patients treated with accelerated partial breast irradiation using 3D conformal radiotherapy. The aim of this study was to analyze the changes in dynamic seroma cavities using fiducial markers in patients with early-stage breast cancer who were treated with APBI using 3D-CRT (three-dimensional conformal external-beam radiotherapy). The investigation evaluated the utility of gold fiducial markers in the aforementioned treatment plan. 34 patients were enrolled in this investigation. Distance changes between the fiducial markers were analyzed and documented. The results showed that there was a strong correlation between the AiMD (average intermarker distance) and seroma volume. The results also showed that exponential reduction in seroma volume impacted the 3D-CRT treatment procedure. It can thus be inferred from this study that AiMD can be used as a surrogate marker for seroma volume¹¹.

In 2010, Yang et al published a study entitled Clinical applicability of cone-beam computed tomography in monitoring seroma volume change during breast irradiation. The aim was to determine the relative effectiveness of main cone-beam CT as compared to conventional CT in monitoring the reduction in the seroma volume during radiotherapy. There were 19 female patients enrolled in this study. All the patients were treated for Stage T1-2 breast cancer, and underwent multiple CT and CBCT scans at specific time intervals during radiation treatment. Seroma contouring was done, and the corresponding seroma clarity and conformity indexes determined. The results showed that there was no significant difference between the CT and CBCT seroma volumes. However, there was a significant correlation between seroma clarity and the CI for both CBCT and CT. It could, therefore, be concluded from this study that the CT and CBCT volume discrepancy was statistically insignificant and that seroma clarity influenced the contouring ability of the observer on either CBCT or CT equally. It can be also be inferred from this study that CBCT can be used as a clinical surrogate for computed tomography during the monitoring of radiotherapy seroma reduction⁸.

Methodology.

The following patients were enrolled in this research:

Patient (1): 48 year old, white female diagnosed with an ER positive high-grade ductal carcinoma in situ of the right breast, status post lumpectomy.

Patient (2): 65 year old white female diagnosed with an ER positive, HER-2 /neu negative, pathologic stage TIcNImiM0 infiltrating ductal carcinoma of the right breast, status post lumpectomy and sentinel lymph node dissection.

Patient (3): 56 year old female diagnosed clinical stage IIa (T2N0M0) infiltrating ductal carcinoma of the right breast. The index lesion was 4 cm in size, ER positive, PR positive.

Patient (4): 58 year old female diagnosed with a pathologic stage 0 grade 3 DCIS (Tis N0M0) of the left breast. The tumor was PR negative and ER positive (4%).

Patient (5): 68 year old female diagnosed with a pathologic stage II (T2N0M0) infiltrating ductal carcinoma of the left breast. The index lesion was 2.3 cm in size, ER positive (greater than 95%), PR positive (75%).

Each of these five patients was subjected to two treatment plans:  initial simulated CT boost treatment plan, and a boost plan ReCT (re-simulated). The Boost Radiation therapy technique used is described below.

To setup the fields, the isocenter was placed on PTV, and the Auto port placed (7mm margin) around PTV using MLC. Six MV beams were used for each field. The plans were then calculated to observe the isodose distributions. Since breasts are conical in shape, the anterior surface of the breast will receive more irradiation. To reduce the hotspots, a physical wedge was inserted in the lateral field. The wedge compensated for the missing tissue and brought the isodose line towards the “toe” of the wedge. The plan was calculated again to see the new isodose distribution. Weighting was applied to distribute the dose uniformly. Normalization was adjusted to increase the PTV coverage.

By looking at DVH

Patient1 (CT)	Total volume	min	max	mean	Cursor volume 1200 cGy
GTV	94.07cc	1195	1346	1264	99.74%
CTV	296.35cc	1185	1362	1275	99.44%
PTV	467.94cc	1116	1365	1278	99.52%

Global max dose 1285 cGy

Hotspot 5%

Patient1 (ReCT)	Total volume	min	max	mean	Cursor volume 1200 cGy
GTV	122.36cc	1187	1323	1238	99.98%
CTV	347.53cc	1335	1242	1242	99.28%
PTV	515.45	1137	1335	1244	98.40%

Global max dose 1289 cGy

Hotspot 5%

Patient2 (CT)	Total volume	min	max	mean	Cursor volume 1200 cGy
GTV	51.65cc	1195	1278	1238	99.99%
CTV	158.84cc	1149	1306	1240	99.39%
PTV	243.19cc	1144	1315	1242	98.39%

Global max dose 1346 cGy

Hotspot 9%

Patient2 (ReCT)	Total volume	min	max	mean	Cursor volume 1200 cGy
GTV	80.46cc	1127	1313	1251	99.24%
CTV	230.87cc	1101	1338	1256	98.90%
PTV	353.57cc	1065	1346	1260	98.89%

Global max dose 1315 cGy

Hotspot 7%

Patient3 (CT)	Total volume	min	max	mean	Cursor volume 1200 cGy
GTV	13.92cc	1159	1269	1243	98.86%
CTV	69.56cc	1123	1284	1235	94.61%
PTV	108.77cc	1118	1286	1232	91.68%

Global max dose 1285 cGy

Hotspot 5%

Patient3 (ReCT)	Total volume	min	max	mean	Cursor volume 1200 cGy
GTV	14.62cc	1165	1281	1249	99.19%
CTV	71.23cc	1089	1286	1239	95.30%
PTV	114.35cc	1080	1289	1234	92.24%

Global max dose 1289 cGy

Hotspot 5%

Patient4 (CT)	Total volume	min	max	mean	Cursor volume 1200 cGy
GTV	60.98cc	1093	1270	1234	96.24%
CTV	151.88cc	1079	1270	1231	93.24%
PTV	202.41cc	1079	1270	1232	92.63%

Global max dose 1270 cGy

Hotspot 5%

Patient4 (ReCT)	Total volume	min	max	mean	Cursor volume 1200 cGy
GTV	25.37cc	1162	1291	1249	99.74%
CTV	91.67cc	1130	1292	1241	97.27%
PTV	154.89cc	1114	1292	1238	95.78%

Global max dose 1291 cGy

Hotspot 6%

Patient5 (CT)	Total volume	min	max	mean	Cursor volume 1200 cGy
GTV	44.39cc	1205	1288	1237	100%
CTV	138.05cc	1119	1280	1242	98.92%
PTV	206.41cc	1061	1281	1240	95.80%

Global max dose 1264 cGy

Hotspot 3%

Patient5 (ReCT)	Total volume	min	max	mean	Cursor volume 1200 cGy
GTV	140.82	1199	1282	1246	100%
CTV	322.90	1131	1287	1245	97.99%
PTV	395.95	1113	1288	1245	97.27%

Global max dose 1254 cGy

Hotspot 2%

DVH analysis was used to evaluate and compare dose among various treatment plans. Data analyzed from the DVH includes: dose coverage of the breast boost volume (GTV, CTV, and PTV) max and mean doses. As well as lung dose (V₅ and V₂₀ of ipsilateral lung and the maximum and mean dose of contralateral lung), and heart dose (V₅ and V₃₀).  The results were collected for both plans. 

References.

1.      Alderliesten, T., den Hollander, S., Yang, T. I. J., Elkhuizen, P. H., van Mourik, A. M., Hurkmans, C., & van Vliet-Vroegindeweij, C. Dosimetric impact of post-operative seroma reduction during radiotherapy after breast-conserving surgery. Radiotherapy and Oncology, 100(2):265-270; 2011.

2.      Chadha, M.; Young, A.; Geraghty, C.; Masino, R.; Harrison, L. Image guidance using 3D-ultrasound (3D-US) for daily positioning of lumpectomy cavity for boost irradiation. Radiation Oncology, 6(1): 45; 2011.

3.      Halperin, E. C.; Perez, C. A.; Brady, L. W. Perez and Brady’s principles and practice of radiation oncology. Elmsford, NY: Wolters Kluwer Health; 2008.

4.      Kader, H. A.; Truong, P. T.; Pai, R.; Panades, M.; Jones, S.; Ansbacher, W.; Olivotto, I. A. When is CT-based postoperative seroma most useful to plan partial breast radiotherapy? Evaluation of clinical factors affecting seroma volume and clarity. International Journal of Radiation Oncology* Biology* Physics, 72(4):1064-1069; 2008.

5.      Kosztyla, R.; Olson, R.; Carolan, H.; Balkwill, S.; Moiseenko, V.; Kwan, W. Evaluation of Dosimetric Consequences of Seroma Contour Variability in Accelerated Partial Breast Irradiation Using a Constructed Representative Seroma Contour. International Journal of Radiation Oncology* Biology* Physics, 84(2):527-532; 2012.

6.      Kumar, V.; Abbas, A.; Fausto, N,; Aster, J. C. Robbins & Cotran pathologic basis of disease. Philadelphia, PA: Elsevier Health Sciences; 2009.

7.      Sharma, R.; Spierer, M.; Mutyala, S.; Thawani, N.; Cohen, H. W.; Hong, L.; Kalnicki, S. Change in seroma volume during whole-breast radiation therapy. International Journal of Radiation Oncology* Biology* Physics, 75(1):89-93; 2009.

8.      Yang, T. I. J.; Minkema, D.; Elkhuizen, P. H.; Heemsbergen, W.; van Mourik, A. M.; van Vliet-Vroegindeweij, C. Clinical applicability of cone-beam computed tomography in monitoring seroma volume change during breast irradiation. International Journal of Radiation Oncology* Biology* Physics, 78(1):119-126; 2010.

9.      Yang, Z.; Chen, J.; Hu, W.; Pan, Z.; Cai, G.; Yu, X.; Guo, X. Planning the breast boost: How accurately do surgical clips represent the CT seroma? Radiotherapy and Oncology, 97(3):530-534; 2010.

10. Yang, Z.; Chen, J.; Xie, J.; Zhu, C.; Pan, Z.; Yu, X.; Guo, X. Simultaneous integrated boost in breast conserving radiotherapy: is replanning necessary following tumor bed change?. Technology in cancer research & treatment, 12(2):115-122; 2013.

11. Yue, N. J.; Haffty, B. G.; Kearney, T.; Kirstein, L.; Chen, S.; Goyal, S. Tracking the dynamic seroma cavity using fiducial markers in patients treated with accelerated partial breast irradiation using 3D conformal radiotherapy. Medical physics, 40: 021717; 2013.