Is there a method to determine the best cardiac-sparing ...

? A retrospective study of left-sided breast patients to determine optimal lung and cardiac-sparing treatment position. Nathaniel Miles, Maddisen Fain, Mary Keating R.T.(T)., Ashley Hunzeker, M.S., CMD, Nishele Lenards, M.S., CMDMedical Dosimetry Program at the University of Wisconsin-La Crosse, WIABSTRACTThe goal of this retrospective study was to evaluate if a tool or metric can be developed that predicts optimal cardiac and lung sparing without the need for patients to undergo multiple CT scans. Thirty-five females diagnosed with breast cancer contained to the left breast were simulated in three different positions: supine free-breathing (FB), supine deep inspiration breath-hold (DIBH), and prone. Treatment plans prescribed to 45 Gy in 28 fractions were created in each position utilizing the Eclipse treatment planning system (TPS). Tangential treatment field border parameters were set as outlined in the Radiation Therapy Oncology Group (RTOG) breast atlas. The heart and ipsilateral lung were contoured as per the RTOG 1106 thoracic atlas and the mean and maximum were compared between each simulation position. Patient factors including BMI, smoking status, breast size, and primary tumor quadrant were evaluated. P-value tests were performed to evaluate the change in mean heart and lung doses. Scatter plots were created to evaluate for correlation between sternal separation and the change in mean heart and lung doses from FB to DIBH scans. It was concluded that although this study failed to find correlation between any measurements that would be predictive of the best position for the treatment of left sided breast cancer patients; however, each patient should be at a minimum simulated in the prone or DIBH position to ensure a reduction in the mean heart dose.Introduction The American Cancer Society estimates greater than 266,000 new cases of breast cancer will be diagnosed in the United States in 2018. The 5-year survival rate is 99% for patients with disease in only one breast without nodal involvement.1 Since most individuals diagnosed with early stage breast cancer will be cured of the disease survivorship is an important topic. A recent review suggested that radiation therapy (RT) should be recommended for 87% of breast cancer patients.2 The main long-term hazards of breast cancer irradiation are that it may cause a second primary cancer in the lung or cardiac complications.3 Minimizing dose to the heart and lung is of vital importance to increase the chance of a morbidity-free survival. The increasing survival rate for breast cancer patients has prompted a growing concern to reduce the risk of developing a secondary lung malignancy.4,5 Grantzau4 et al evaluated the effects of breast cancer radiation to the risk of developing a second primary lung cancer and found an 8.5% linear increase per Gy. Zhao et al6 showed that DIBH can achieve lower ipsilateral lung and cardiac doses compared to FB plans. Lymberis et al7 evaluation of prone versus supine positions found the prone setup reduced the amount of irradiated lung in all patients. The study also found prone positioning to reduce the amount of heart volume irradiated in 87% of left breast patients. These studies show that utilization of either prone or DIBH position can lead to a reduction in morbidity. However, smoking can decrease chest wall expansion, forced vital capacity, and maximal expiratory muscle strength.8 Neugut et al5 found a 30-fold increased risk of developing a second primary cancer in the ipsilateral lung of breast cancer patients who smoked cigarettes while undergoing RT. It is unclear if the reduction in chest wall expansion correlates with the optimal position for left –breast patients who smoke.In left breast RT treatments, the heart can receive increased exposure to ionizing radiation due to its location when compared to right-sided tumors. Increased exposure can increase the subsequent rate of ischemic heart disease, which has been found to be proportional to the mean heart dose.4 Additionally, the rates of major coronary events increase linearly with the mean heart dose by 7.4% per Gy.4 A recent international breast cancer randomized trial (NSABP B-51/RTOG1304) proposed the mean heart dose (Dmean Heart) should be less than 4 Gy during left-sided breast/chest wall irradiation.5 Several studies have been performed that acknowledge the benefit of utilizing different positioning and treatment techniques to reduce cardiac dose.2,5,7,8,9,10,14 Yet, there are a lack of guidelines as to the best practice in cardiac sparing.7DIBH and prone positioning RT planning techniques are currently recommended by the National Comprehensive Cancer Network (NCCN) to reduce heart and lung dose.6 The widespread availability of prone breast boards and surface-guided RT systems have led to radiation oncology departments being able to perform multi-positional CT simulations. No tool or metric has been published that allows for evaluation of best scenario heart and lung sparing in left-sided breast patients without having to obtain CT scans in numerous treatment positions. Chen et al8 evaluated 25 women by measuring the heart to tangent field border looking for correlation to cardiac dose reduction but it only compared heart-sparing in the prone versus supine positions. Walston et al9 evaluated the dosimetric sparing of organs using the DIBH technique versus FB and sought to find a correlation between the chest wall excursion from the couch top and benefits in heart and lung sparing. Although Walston et al9 failed to find a correlation, they revealed a dosimetric advantage could be better achieved by measuring heart to tangential field border. Nissen and Appelt10 also evaluated FB versus DIBH and found there to be a lack of predicting factors for whether a patient will benefit substantially from DIBH. For individual patients it is difficult to predict the optimal treatment position without taking both FB and DIBH CT scans into consideration. The existing literature showcased the need for the current study to be conducted. This study compared patients in 3 positions to evaluate if a tool or metric can be developed that predicts optimal cardiac and lung sparing without the need for patients to undergo multiple CT scans. Individuals diagnosed with early stage left-sided breast cancer a high rate of success for cure. Therefore, it is imperative that dosimetric measures be taken to ensure that their survivorship is not burdened by preventable co-morbidities. Additionally, in adherence with the radiation principle of “as low as reasonably achievable” (ALARA), radiation oncology professionals should strive to find a guideline for a best practice in cardiac sparing and lung dose reduction. The goal of this retrospective study was to develop a tool that predetermines best position for optimal heart and lung sparing to minimize multiple CT scans.Methods and Materials Patient Selection Thirty five patients selected for this study were diagnosed with cancer of the left breast without lymph node involvement. Each patient underwent CT simulations in the prone, supine FB, and supine DIBH positions. The simulations were performed in 1 of 4 centers within a multi-site network. In all simulations, the field borders were delineated with wire. The scan length included the entirety of the ipsilateral lung with the superior border at the level of the chin and the inferior border 5 cm below the visible breast tissue.During simulation, each patient underwent a FB and DIBH CT scan on a MedTec breast board with the arms overhead (Figure 1). A Vacloc bag was positioned under the torso for immobilization purposes. The tumor bed scar, field borders, nipple, and cental axis (CAX) were delineated with radio opaque wire or markers (Figure 2). The chin was turned to the contralateral side and raised slightly. Once properly immobilized, the patient underwent both FB and DIBH CT scans using either a GE Lightspeed or a Philips Brilliance Big Bore CT scanner with 2.5 mm axial slice thickness. First, the FB scan was performed. The breath hold was timed for 20 seconds at which the patient was then directed to breathe. This was repeated at least once to ensure the patient was comfortable with the process and duration of the breath hold prior to the scan. Upon demonstration of appropriate breath hold technique the patient was scanned holding their breath. Following completion of the FB and DIBH simulations, the patient was simulated in the prone position on a CDR prone breast board. The radiation therapist indexed the prone breast board to ensure that the hole for breast placement was large enough to keep the field borders from touching the breast board whenever possible. The midline marker was kept 1-2cm from the breast board whenever possible. Care was taken to ensure no skin folds existed between midline and the contralateral breast and inferior breast tissue and abdomen. The head was turned in the direction that the patient found most comfortable. The borders from supine position were used. A radio opaque marker designating the prone position CAX was placed on the outer-laterality, midway between the superior and inferior borders, of the left breast (Figure 3). The patient was then scanned in the prone position. Each scan was then sent to Raystation TPS. ContouringThe patient DICOM data sets were anonymized and sent to 2 centers for contouring. Contouring was performed on each of the prone, DIBH, and FB using either Raystation or Eclipse TPS. The contours included the heart, ipsilateral lung, and the radio opaque markers which indicated the location of the nipple and each of the field borders drawn by the physician at the time of CT simulation (Figure 4). The physician designated field borders to comply with the RTOG Breast Atlas. The superior field border was placed at the level of the second rib insertion; the caudal border was placed inferior to the inframammary fold using clinical judgement to delineate the termination of breast tissue. The lateral border was placed at the mid axillary line, excluding the latissimus dorsi. Lastly the medial border was placed midline at the level of the sternal rib junction.11The heart and left lung were contoured referencing the RTOG 1106 Thoracic Atlas. The lungs were contoured using pulmonary windows. All inflated and collapsed, fibrotic and emphysemic lung tissue was contoured; the small vessels extending beyond the hilar region were included. The hila and trachea were not included in the lung tissue contour. The heart and pericardial sac were contoured. The superior aspect of the heart began at the level of the inferior aspect of the pulmonary artery passing the midline and extended inferiorly to the apex.12 Treatment PlanningTangential fields were created utilizing the borders indicated by the radiation oncologist at the time of simulation. The medial radio opaque marker was used as a user origin/localization point for all scans. The same isocenter shift was used for the FB and DIBH CT scans as the medial marker was not moved between scans. The DIBH shift was found first by placing the isocenter in the center of the breast. The prone isocenter was placed in the center of the breast. The gantry was rotated until the medial and lateral marker were in alignment. The collimator was rotated to follow the chest wall. A rectangular field was created using the field border defining markers placed at the time of CT simulation and 2cm of flash was created anteriorly (Figure 5). Medial and lateral tangents were created, and field divergence was matched along posterior border. All plans were calculated to receive 45 Gy in 25 fractions at 180 cGy per day. The plans were calculated using 6 MV energy if the posterior field border on the localization slice was less than 24 cm from medial to lateral. If the border was larger than 24 cm, 18 MV was used. The DIBH plan was normalized first by placing a calculation point in the same axial plane as the isocenter 1 cm from the chest wall halfway between the medial and lateral borders of the posterior field edge. The DIBH plan was then normalized to this point. All other plans used the same normalization values as the DIBH plan. All plans were calculated for the same Varian TrueBeam linear accelerator using Eclipse TPS and Analytical Anisotropic Algorithm (AAA). To maintain consistency between the plans, none were weighted, no blocks were drawn, no field-in-fields were used, and no wedges were added. The plans were kept as open fields with equal beam weighting. Once the fields were created and dose was calculated, both the mean and maximum doses for the heart and lung were measured for all patients in each of the 3 positions. Additionally, the hotspot was recorded for each plan. A sternal separation measurement was obtained for the DIBH by measuring the difference in distance from couch top to skin between DIBH and FB (Figure 6). A heart separation from tangential field to pericardium between DIBH and FB was also obtained (Figure 7). The breast size was recorded as well and was determined by measuring from the medial to lateral field edge at the level of isocenter (Figure 8).Plan ComparisonsThe mean and maximum heart and lung doses were averaged for each of the FB, DIBH, and prone positions. P-value tests were performed to evaluate the change in mean heart and lung doses. P-values of ≤ 0.05 were considered statistically significant. The mean heart and lung doses with each technique were then compared to several different patient measurements to evaluate for correlation. The mean heart and lung dose were compared to the breast size, breast quadrant, smoking status and BMI. A scatter plot was created to determine if there was a correlation between heart separation and the change in mean heart dose from FB to DIBH scans. A comparison between heart separation and the change in lung/heart dose from FB to DIBH scans was also made using a scatter plot. Additional scatter plots were created to evaluate for correlation between sternal separation and the change in mean heart and lung doses from FB to DIBH scans.A table of demographic information and the change in mean lung and heart dose associated with the different groups was created to see if specific techniques were superior based on demographics. Some of the demographics evaluated included whether a patient was a current or former smoker, whether a patient was considered obese from their BMI, a patient's breast size (less or greater than 24cm), and the seroma location within the breast quadrant.ResultsThe study included 35 females with left sided breast cancer. Table 1 represents the clinical characterization of the patients involved in the study. As shown, the patients' age ranged from 30 to 75?years with a median of 58 years. The median BMI was 28.7. Only 1 woman was a current smoker, 16 women were former smokers, and 18 have never smoked. Most patients were diagnosed with a type of ductal carcinoma, with 34.3% having ductal carcinoma in situ (DCIS) and 48.6% having invasive ductal carcinoma (IDC). Only 2.8% of patients had lobular carcinoma in situ (LCIS) and 14.3% were diagnosed with invasive lobular carcinoma (ILC). Primary tumor location included the following: 45.7% of all tumors were found in the upper outer quadrant (UOQ), 17.1% were located in the upper inner quadrant (UIQ), 14.3% were located in the lower outer quadrant (LOQ) and 22.9% in the lower inner quadrant (LIQ). Only 2.8% of patients had a triple negative immunohistochemistry (IHC) expression of estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (Her2). 42.8% of patients were ER/PR+, Her2- and 14.2% were ER/PR+, Her2+. The DCIS and LCIS patients were not tested for Her2, but all were found to be ER/PR+. Only 5.6% of patients were found to be ER/PR- (Table 1).The average mean and maximum heart dose were found to be higher in the FB plan than in DIBH or the prone plans (Table 2). The average mean heart dose was 182cGy for the FB plan versus 132cGy for the DIBH plan and 120cGy for the prone plan. The average maximum heart dose was 2798cGy for the FB plan versus 1523cGy for the DIBH plan and 1739cGy for the prone plan. The mean and maximum lung doses were highest in the DIBH plans with a mean lung dose of 524cGy and the maximum lung dose was 3858cGy. The average mean DIBH lung dose was 0.8% higher than the FB plan and10.6% higher than the prone plan. The average maximum DIBH lung dose is 2.6% higher than the FB plan and 58.1% higher than the prone plan. P-value tests were performed to evaluate the change in mean heart and lung doses (Tables 3-4). The change in mean heart dose between positions was found to be statistically significant for the FB versus DIBH and prone positions. There was no statistical significance between the prone and DIBH positions. There was also a statistical significance in the change of mean lung dose for the prone position as compared to both the FB and DIBH positions. The statistics showed that overall DIBH was better for mean and maximum heart dose with an average of 2.93% and 33.84% of prescription, respectively, as compared to 4.04% and 62.17% for the FB scans. Prone proved to be the best option dosimetrically for virtually all the patients, with improvements in mean and maximum heart and lung dose in comparison to both the FB and DIBH scans.The dose statistics between the factors that were assessed, such as breast size, sternal separation, smoking status, and breast quadrant failed to show any correlation (Table 5). The heart separation versus FB-DIBH mean heart dose did show some correlation when the heart separation measured was over 1.5cm (Figure 9-12). This showed that patients with a large amount of heart separation, at least in the plane measured, had the greatest benefit from a DIBH treatment plan. Patients with less heart separation only averaged around 1% mean heart dose sparing while patients over 1.5cm of separation showed increasing benefit, with the patient who had a heart separation of around 2.5cm of separation receiving the most benefit of over 5% improvement. DiscussionThis study failed to find correlation between any measurements that would be predictive of the best position for the treatment of left-sided breast cancer patients without having to undergo a CT simulation in multiple positions for final determination. The patients' smoking status, BMI, primary tumor quadrant, and breast size were not predictive of mean heart or lung dose. Single slice data obtained at the level of isocenter including sternal separation, heart separation, and breast size were also not predictive of a guideline for optimal positioning practice.A multi-institutional study of greater than 10-year survivors of breast cancer noted that smoking and RT together increased the risk of fatal myocardial infarction (MI).13 Smoking has also been found to increase the risk of developing a second primary lung malignancy especially in patients undergoing breast RT.# Unfortunately, in the current study, patient smoking status was not indicative of the mean heart or lung dose. The smoking status did not correlate to the sternal or heart separation. A measurement of lung volume might have correlated better to sternal or heart separation. Due to the increased risk of MI and second lung primary, left-sided breast cancer patients who maintain a current smoker status should be offered simulations in each position to determine which is most advantageous. It is known that the DIBH technique reduces the mean heart dose through immobilization of the chest wall and by increasing the distance from the heart.14 However, sternal separation measured at the level of isocenter did not prove to be a good indicator of improvement between the DIBH and FB positions. This could simply be attributed to the fact that all patients breathe differently, with some breathing more from their diaphragm and others breathing more with their whole chest.15 Furthermore, in the prone position where sternal separation was negligible, the mean heart dose was found to be lower than that of both FB and DIBH plans. The measurement of heart separation at the level of isocenter did not show correlation to the mean heart or lung dose. This might be attributed to the position of the heart at the level of isocenter. Greater separation might be seen at another level of the heart such as the apex and a distance measurement could potentially prove to be statistically significant. Additionally, the heart moves with the respiratory cycle so the degree of motion, mainly in the superior–inferior direction, is modest with free breathing.15 An ECG may be a way to minimize radiation exposure and determine the proximity of the heart to the chest wall in various treatment positions.Breast size measured on a single slice and BMI showed no correlation to each other. They also were not predictive of heart and lung doses. However, Brown et al16 has shown that BMI and suprasternal notch to nipple distance enable predictions of breast mass. A measurement of breast mass such as this may show correlation to lung and heart dose.Finally, although primary tumor breast quadrant statistics did not correlate to lung or heart dose in this study, the quadrant location may be predictive of positional benefits when a tumor bed boost is prescribed. The patients in this study did not have a boost prescribed. Positioning a patient based on the cavity might reduce the heart and lung from additional exposure as it relates to the boost.ConclusionIn conclusion, this study failed to find correlation between any measurements that would be predictive of the best position for the treatment of left sided breast cancer patients without having to undergo a CT simulation for final determination. Breast size, breast quadrant, smoking status, and BMI did not prove to be measures that correlate to the best treatment position. However, because the prone and DIBH position showed a significant decrease in the mean heart dose as compared to the FB scans, patients should be offered treatment in one of these positions if they are able to tolerate it. One possible limitation of this study was the limited sample size used. This retrospective analysis included only 1 current smoker. If a larger number of patients with a current smoking habit were included in the study then a correlation between heart and lung dose may have been discovered. The more comprehensive data could then be used to evaluate treatment positions. Another limitation of this study was the evaluation of measurements taken in a single plane. The heart separation may have been a more meaningful statistic if measurements were made in each plane and these statistics were averaged. The volume and relative movement of heart determined by an ECG may be more predictive of which treatment position would most likely reduce cardiac dose. Evaluation of breast mass might have proven a more useful statistic than the single measurement of medial to lateral border at the level of isocenter. Centers should continue to perform CT simulations in each of the DIBH, FB, and prone positions to ensure that each patient has the best opportunity for the creation of an individualized treatment plan which best minimizes heart and lung dose during the treatment of left-sided breast cancer.References American Cancer Society. Breast. American Cancer Society Website. July 1, 2018.?Bruzzaniti?V, Abate A, Pinnarò P, et al.?Dosimetric?and clinical advantages of deep inspiration breath-hold (DIBH) during radiotherapy of breast cancer.?J Exp?Clin Cancer Res.?2013;32(1):88-94.? C, Correa C, Duane FK, et al. Estimating the risks of breast cancer radiotherapy: evidence from modern radiation doses to the lung and heart from previous randomized trials. J?Clin?Oncol.?2017;35(15):1641-1649.? SC,?Ewertz M, McGale P, et al.?Risk of ischemic heart disease in women after radiotherapy for breast cancer. N Engl J Med. 2013;368(11):987-98. , Conroy L, Long K, et al. Cardiac dose reduction with deep inspiration breath hold for left-sided breast cancer radiotherapy patients with and without regional nodal irradiation.?Radiat Oncol.?2015;10:200-206.? Comprehensive Cancer Network. NCCN Guidelines Version 1.2018 Breast Cancer. National Comprehensive Cancer Website. July 1, 2018.?Duma MN, Münch S,?Oechsner?M, Combs S. Heart-sparing radiotherapy in patients with breast cancer: what are the techniques used in the clinical routine?: A pattern of practice survey in the German-speaking countries. Med Dosim. 2017;42(3):197-202. JL, Cheng JC,?Kuo?SH, Chan HM, Huang YS, Chen YH. Prone breast forward intensity-modulated radiotherapy for Asian women with early left breast cancer: factors for cardiac sparing and clinical outcomes.?J?Radiat?Res. 2013;54(13):899-908. S, Quick AM, Kuhn K,?Rong?Y.?Dosimetric?considerations in respiratory-gated deep inspiration breath-hold for left breast irradiation.?Technol?Cancer Res Treat. 2017;16(1):22-32.? Nissen HD, Appelt AL. Improved heart, lung and target dose with deep inspiration breath hold in a large clinical series of breast cancer patients.?Radiother?Oncol. 2013;106(1):28-32.? White?J, Tai A, Arthur D, et al.?Breast Cancer Atlas for Radiation?Therapy Planning:?Consensus?Definitions.?Radiation?Therapy Oncology Group website. 1, 2018.?Kong FM, Ritter T, Quint DJ, et al.?Consideration of dose limits for organs at risk of thoracic radiotherapy: atlas for lung, proximal bronchial tree, esophagus, spinal cord, ribs, and brachial plexus.?Int?J?Radiat?Oncol?Biol?Phys.?2011;81(5):1442-1457.? MJ, Botma A, Aleman BM, et al. Long-term risk of cardiovascular disease in 10-year survivors of breast cancer. J Natl Cancer Inst. 2007;99(5):365-375. HB, Duffy D, Yamoah K, et al. Modeled risk of ischemic heart disease following left breast irradiation with deep inspiration breath hold. Pract Radiat Oncol. 2015;5(3):162-168. K, Mead J. Measurement of the separate volume changes of rib cage and abdomen during breathing. J Appl Physiol.1967;22(3):407-422. R, Moran JM, Kessler ML, Marsh RB, Balter JM, Pierce LJ. Respiratory motion of the heart and positional reproducibility under active breathing control. Int J Radiat Oncol Biol Phys. 2007;68(1):253-258. N, White J, Milligan A, et al. The relationship between breast size and anthropometric characteristics. Am J Hum Biol. 2012;24(2):158-164. FiguresFigure 1. Patient setup lying supine on the breast board from MedTec with the arms positioned overhead.Figure 2. Patient markings in the supine position.Figure 3. Patient setup lying prone on the breast board from CDR with the CAX indicated by the mid-breast BB.Figure 4. Radio opaque markers (red) placed on the nipple and at the level of the field borders drawn at the time of CT simulation.Figure 5. Tangential field creation and collimator rotation set by following the chest wall.Figure 6. The sternal separation measured on both the DIBH (left) and FB (right) CT data sets.Figure 7. The heart separation measured on both the DIBH (left) and FB (right) CT data sets.Figure 8. The breast size measurement as determined by measuring from the medial to lateral field edge at the level of isocenter. Figure?9.?A comparison between heart separation and the change in mean heart dose from FB to DIBH scans.?Figure?10.?A comparison between heart separation and the change in mean lung dose from DIBH to FB scans.?Figure?11.?A comparison between sternal separation and the change in mean heart dose from FB to DIBH scans.?Figure?12.?A comparison between sternal separation and the change in mean lung dose from DIBH to FB scans.?TablesTable 1.?The clinical characterization of patients.Parameter TotalAge (Years)Min-MaxMedian30-7558Weight (Lbs.)Min-MaxMedian115-246173Height (in.)Min-MaxMedian156-6965BMI Min-MaxMedian22-42.628.7Smoking Status (%)NonFormerCurrent51.445.82.8Tumor Quadrant (%)UOQLOQUIQLIQ45.714.317.122.9Diagnosis (%)DCISInvasive DuctalLCISInvasive Lobular34.348.62.814.3IHC Expression (%)ER/PR+, Her2+ER/PR+, Her2-ER/PR+, Her2 unknownER/PR-, Her2+ER/PR-, Her2-14.242.837.42.82.8Table 2. Average mean and maximum dose for heart and lung for each scan.Parameter?FB?(%)DIBH?(%)Prone?(%)Mean heart?4.04?2.93?2.66?Max heart?62.17?33.84?38.64?Mean lung?10.8?11.64?1.08?Max lung?85.29?87.95?29.88?Table 3.?Statistical significance of change in mean heart dose (n=35).?Parameter?P-Value?FB versus DIBH?0.001?FB versus Prone?0.033?DIBH versus Prone?0.54?Table 4.?Statistical significance of change in mean lung dose (n=35).?Parameter?P-Value?FB versus DIBH?0.13?FB versus Prone?0?DIBH versus Prone?0?Table 5. Average change in mean heart dose between scans for patient demographic data. Parameter?FB-DIBH?(%)FB-Prone?(%)DIBH-Prone?(%)Non-smoker?1.31?1.75?0.44?Current smoker??1.1?2.5?1.4?Former smoker?0.89?0.91?0.01?BMI >30%?1.06?1.23?1.72?BMI <30%?1.16?1.54?0.38?UOQ?Quadrant??1.61?2.41?0.79?LOQ?Quadrant??0.32?-0.04?-0.36?UIQ?Quadrant??1.2?1.17?-0.03?LIQ?Quadrant??0.44?0.21?-0.23?Breast size >24cm?1.04?1.44?0.4?Breast size <24cm?1.36?1.13?-0.23? ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download