Breast Edema following Staging Axillary Node Dissection in Patients with Breast Carcinoma Treated by

Radical Radiotherapy

DANIEL CLARKE, MD,' ALVARO MARTINEZ, MD,t RICHARD S. COX, PHD,S AND DON R. GOFFINET, MDS

Seventy-four patients with carcinoma of the breast were treated by irradiation without a mastectomy at Stanford between May 1973 and March 1980. Seventy-six breasts were treated because two patients had bilateral simultaneous cancers. Breast edema was noted in 41% of these cases. Analysis of potential predisposing factors revealed that this complication occurred primarily in patients who received a staging axillary lymph node dissection. Lymphedema occurred in 26 of 33 breasts (79%) in patients who had staging uxillary dissections; this complication developed in only three of 12 (25%) with axillary samplings and two of 31 (6%) with no axillary surgery. Two groups of patients were identified: (1) 28 patients whose breast edema occurrred early (prior to or during radiation therapy), and (2) three patients with late onset edema that developed several months postirradiation. The early edema, which is clearly related to axillary lymph node dissection, gradually improved during the follow-up period with complete or near complete resolution expected by three years. Late onset edema was rare, appears to be related to an axillary radiation dose greater than 5500 rad, and may be irreversible. There was no correlation of breast edema with tumor stage, radiation time/dose factors, the use of bolus, patient weight, or breast size.

Cancer 49:2295-2299, 1982.

t iE ROLE OF RADIOTHERAPY in the primary treat- T ment of early stage breast carcinoma is well es- tablished. Local control and survival (at five and ten years) for several large radiotherapy series' compare favorably with results following radical m a s t e c t ~ m y . ~ . ~ Local control rates in excess of 90% with current ra- diotherapy techniques have been reported in Stage I and I1 cancers.6 l o Because the major advantage of ra- diation therapy is breast preservation, attention to fac-

Presented at the 22nd Annual Meeting of the American Society of Therapeutic Radiologists, Dallas, Texas, October 21 -25, 1980.

From the Department of Radiology, Paul A. Bissinger Memorial Center for Radiation Therapy, Stanford University School of Medi- cine, Stanford, California.

Supported in part by Grant CA-05838-19 from the National Cancer Institute, National Institutes of Health, Bethesda, Maryland.

Resident, Division of Radiation Therapy. + Assistant Professor, Division of Radiation Therapy. $ Radiation Physicist. Division of Radiation Therapy. § Associate Professor. Division of Radiation Therapy. Address for reprints: Alvaro Martinez, MD, Department of On-

cology, Division of Radiation Therapy, Mayo Clinic, Rochester, M N 55905.

The authors thank our brachytherapy technicians, Donna Pooler and Mary Jane Schulz, (or their expertise, as well as all the Radiation Therapy technologists and residents whose skill and clinical excellence made this study possible. We also thank Maria Melendez who pre- pared the manuscript.

Accepted for publication March 23, 198 I .

tors that maximize the cosmetic result is becoming more important. For example, the type of biopsy incision (along skin lines and circumareolar when possible), the extent of the biopsy procedure, radiation time/dose l'ac- tors, and careful radiation technique have been shown to be of significant ' * I 2

One complication that has not been previously ana- lyzed and is frequently observed in this group of pa- tients, is lymphedema of the breast. This study attempts to clarify the etiology and natural history of breast edema by analyzing its relationship to treatment vari- ables and other predisposing factors. The severity of breast edema, time of onset, and degree of resolution are discussed. Our treatment technique and results are presented.

Materials and Methods

Patient Population

Seventy-four patients (median age 46) with biopsy- proven adenocarcinoma of the breast were treated by radiation without a mastectomy at Stanford between May 1973 and March 1980. Seventy-six breasts were treated because two patients had simultaneous bilateral cancers. The median follow-up was 23 months (range

0008-543)</82/0601/2295 $0.75 Q American Cancer Society

2295

2296 CANCER June I 1982 Val. 49

FIG. I . A 44-year-old premenopausal woman with Stage I infil- trating ductal carcinoma of the right breast. Mild edema preceded radiotherapy. Photograph was taken during treatment.

12 months to 7.5 years). All patients were clinically staged according to current UICC-TNM ~ r i t e r i a . ' ~

Treatment Techniques

All patients were treated with a combination of 4 MeV x-rays to the involved breast and regional nodes followed by a boost to the tumor bed (67 by '921r im- plant, five by 9-15 MeV, four by 4 MeV x-rays). The internal mammary (IM), supraclavicular (SCV), and axillary (AX) regions were treated in continuity by a single anterior field angled 15-20' laterally to avoid irradiating the vertebral bodies, spinal cord, and esoph- agus. The IM portion extended medially 1 cm across the midline, was 5-cm wide, and encompassed the first 3-4 intercostal spaces. The field extended superiorly to the cricothyroid groove and laterally to 2 cm inside the pectoralis major muscle (unless bulky axillary adeno-

FIG. 2. A 41-year-old premenopausal woman with simultaneous iti siru carcinoma of the right breast and infiltrating ductal carcinoma of the left breast. who underwent left axillary lymph node dissection. Both hreasts were treated with identicalexternal beam radiation doses. followed by bilateral 192-Iridium implants. Photograph illustrating moderate left breast edema was taken at first follow-up.

pathy was present) to avoid moist desquamation and muscle fibrosis. Coverage of the interpectoral nodes was assured by the tilt, and the humeral head was excluded by a shaped lead block. The inferior edge of the SCV/ AX portion was level with the second intercostal space. Internal mammary and SCV doses were calculated to a depth of 3 cm, AX doses to the midplane. A posterior- axillary field was used during the final third of treat- ment to supplement the axillary dose as required.

The breast and chest wall were treated by tangential fields, usually parallel opposed. The field was blocked at its central axis with a beam splitter, thus decreasing beam divergence and minimizing the volume of irra- diated lung.14 The superior and medial borders abutted the anterior field. The inferior border was at the infra- mammary fold and the lateral border at the mid or posterior axillary line depending on the posterior extent of dependent breast tissue. Wedges, or when indicated, individually constructed compensators that correct for surface irregularity, were employed to provide a more homogeneous dose distribution. Bolus to the chest wall was not used unless the skin was involved by carcinoma.

Radiation doses and fractionation schedules evolved during the six-year interval. Initially 250 rad per frac- tion, four times a week, to a minimum tumor dose of 5000 rad to the breast and regional nodes was given to six patients. Later the dose per fraction was reduced to 200 rad and given five times a week for a dose of 5000 rad. Eighteen patients were treated in this fashion.

Our current policy is to deliver a minimum tumor dose of 4500 rad at 180 rad per fraction, five times per week. Dose adjustments were based on the status of the axillary nodes, quadrant location of the tumor, and the adequacy of the tumor biopsy. I n patients receiving an axillary dissection, the lower two-thirds of the axilla was not treated unless there was extranodal extension of tumor. Three to four weeks following external beam radiation, a I9*Ir implant was performed as previously described,' delivering an additional 2000-2500 rad to the tumor volume.

Axillary Staging

Premenopausal women, who were candidates for ad- juvant chemotherapy, generally received a staging ax- illary lymph node dissection prior to radiation therapy. This procedure, well standardized at our institution, consisted of identifying the axillary vein and resecting the axillary fat and lymph nodes inferior to it, with an average of 18 nodes removed. Thirty-three patients had dissections. Because of surgeon preference twelve pa- tients had axillary samplings, with 1-6 nodes removed. Thirty-one patients had no axillary surgery. Most were postmenopausal, although some premenopausal pa- tients refused axillary surgery.

No. I I LYMPHEDEMA OF T H E BREAST * Clarke et 01. 2297

Assessment and Analysis of Complications

At each follow-up visit, in addition to the routine clinical examinations, an evaluation of various treat- ment complications was meticulously performed. Breast edema, breast fibrosis, hyperpigmentation, arm edema, superficial thrombophlebitis, and various other compli- cations were scored as mild, moderate, or severe by three of the four authors (A.M., D.C., and D.G.) to insure consistency in evaluation. Data for complications other than breast edema are currently being analyzed for future publication.

Mild breast edema (Fig. 1 ) was asymptomatic and caused only a minor increase in breast size. Moderate edema (Fig. 2) was characterized by a readily apparent disparity between breasts, and was usually accompanied by some degree of skin edema with features of peau d'orange (Fig. 3) . Such patients often complained of subjective breast heaviness. Severe edema was defined as that which caused significant symptoms such as pain or resulted in a poor cosmetic result.

In our analysis of breast edema as a complication of treatment, these covariates were considered as potential risk factors: patient weight, breast cup size, axillary lymph node surgical staging procedure, use of bolus, fractionation scheme, radiation dose to the breast and axilla, axillary lymph node status, and T stage of the tumor. Each parameter was studied independently for its effect on the incidence of breast edema. Contingency tables were prepared and the significance of differences assessed by means of the chi-square test. Those factors showing a correlation with breast edema ( P 5 0.10) were subsequently analyzed according to a multivariate regression method. ' The calculations were performed on a PDP-11-50 computer.

Results

Local Control

In Stages I and 11, local or regional recurrences have developed in two of 66 treated breasts, resulting in a local control rate of 97%. One patient with a Stage I carcinoma presented with a cytologically confirmed malignant nipple discharge. She had a recurrence 35 months postirradiation and has subsequently undergone a radical mastectotny and is receiving chemotherapy. A second patient with Stage I1 carcinoma had axillary adenopathy on her first follow-up visit. Recurrent or persistent disease was pathologically confirmed in both her breast and axilla. She has undergone a local excision and is currently receiving chemotherapy.

Six of ten Stage 111 and IV patients have failed dis- tantly and five are dead of disease. Two patients have had a local recurrence in conjunction with metastatic disease.

FIG. 3. A 42-year-old premenopausal woman with Stage I 1 infil- trating ductal carcinoma of the right breast. Moderate breast edema occurred after axillary dissection. Characteristic peau d'orange was noted.

Lymphedema of the Breast

This complication was observed in 41% of the patient population, and was related to the axillary staging pro- cedure (Table 1). Thirty-three of the 76 carcinomas were staged with axillary dissections; lymphedema of the breast was observed in 26 (79%). Breast edeina occurred in only three of 12 patients who had axillary

TABLE I . Correlation of Axillary Staging Procedure and Breast Edema

Axillary staging

Dissection Sampling None (%) Breast edema (%)

Positive 31 26 (78) 3 (25) 2 (6) Negative 45 7 (22) 9 (75) 29 (94)

- ~ .~

TOTAL 33 (loo) 12 ( l o o ) 31 (100)

An increased incidence of breast edema was observed in patients who received axillary dissections ( P < 0.0001).

TABLE 2. Prognostic Significance of Covariates for Breast Edema ~~

Analysis

Univariate* M ultivariatet

~-

- Covariates P P

Weight 0.94 - Breast size 0.78 -

Fractionation 0.52 - 0.48 - N stage

Axillary dose 0.47 0.10 Breast dose 0.04 0.45 Bolus 0.01 0.41 T stage 0.0 I 0.49

- .-

Axillary surgery < I 0 - 4 < 1 0 - 4 ~ .

* Univariate analysis with P values obtained by chi-square test. I' Multivariate analysis with P values obtained by multivariate

Only axillary surgery was significant. regression method.

2298

- I -

NO DISSECTION

- -

- D ISSECT1 ON

P < 0.0001 I I I I

CANCER June 1 1982 V O l . 49

a 3 100

3

0 w

80

K

I- 60

I I I I 1 I I 1

a E o

0 1 2 3 4 5 6 7 DATE OF DIAGNOSIS TO TIME OF ONSET (years)

FIG. 4. Time of onset of breast edema from date of diagnosis. All 76 treated breasts were considered at risk.

sampling procedures (25%), and in only two of the 31 treated breasts with no axillary surgery ( 6 % ) .

The occurrence of breast edema was not correlated with T stage, axillary nodal status, fractionation scheme, use of bolus, radiation dose to the breast, weight, or breast size (Table 2).

Figure 4 illustrates the pattern of onset of breast edema after the date of diagnosis. All 76 treated breasts were considered at risk. Most cases of breast edema developed within the first one or two months after di- agnosis, although the onset of edema was occasionally noted after 18 months. Ten of 31 cases occurred prior to the initiation of any radiotherapy; 18 cases were first observed during external beam therapy; late onset edema was observed in only three patients (Table 3).

Figure 5 shows the pattern of onset of breast edema as related to surgical axillary staging. All patients with axillary dissections in whom breast edema developed

a I I I I I I I I

did so early (before or during radiotherapy). These pa- tients generally had an abrupt onset of mild edema, the severity of which was occasionally exacerbated by the Iridium implant (four patients) or by additional com- plications (upper extremity superficial thrombophlebitis in four patients) as shown in Table 3. An additional patient developed an exacerbation at the time of a di- agnostic thoracotomy. None of these patients has ex- perienced severe edema. Breast edema that occurred in the nondissected population was uncommon and tended to have a late onset.

For the 28 patients with early breast edema, the time from onset of edema to first improvement (from mod- erate to mild, or mild to complete resolution) has been plotted in actuarial form in Figure 6 (broken line). No patient was at risk for more than two years without demonstrating some degree of resolution. The solid line shows the time to complete resolution of breast edema. Nine of the 28 patients have had the breast edema dis- appear (range 4-29 months). The time to median res- olution was 24 months, with complete resolution ex- pected by three years.

Breast edema developed in three patients following completion of treatment. All had clinically positive ax- illary nodes and received high radiation doses to the axilla. None received an axillary dissection. Breast edema developed in one patient following 6000 rad to the axilla 35 months after presentation; this patient also had slowly progressive edema over the ensuing 18 months, at which time she died of myocardial infarc- tion. An autopsy revealed no evidence of tumor. A sec- ond patient, who received 5500 rad to the axilla and a 6-cm orthovoltage boost of 1500 rad, had persistent breast edema at 20 months. The third patient also re- ceived 5500 rad to the axilla; edema occurred one month postirradiation but resolved completely after 12 months.

Discussion

Although the primary objective of the radiotherapist is to obtain local control of the neoplasm and cure the patient, the importance of obtaining a satisfactory cos- metic result should not be overlooked. Although our data are preliminary, the local control rate of 97% in Stage I and I1 patients in this series is acceptable and supports the results published by other radiotherapy centers.

Lymphedema of the breast that is observed after the primary radiotherapeutic management of breast car- cinoma has often been considered a complication of external beam radiation. Our data, however, indicate that this complication arises primarily from the axillary staging procedure. Breast edema developed in 79% of the patients who underwent axillary lymph node dis- sections as opposed to 12% of patients who had axillary samplings or no axillary surgery at all. This finding

No. I I LYMPHEDEMA OF T H E BREAST Clarke et a / . 2299

I I I

I I I II I 1 I -

clearly demonstrates the importance of the axillary route of breast lymphatic drainage.

Many careful anatomic and dynamic studies of the lymphatic drainage of the breast have been per- formedI6 " and have consistently shown that the axil- lary route of drainage is considerably more important than the internal mammary system. Radioactive l9*Au studies indicate that 75-95% of the breast lymphatic flow is to the axilla."~'' Such data are supported by well-known clinical observations that lymph node me- tastases from breast carcinomas are far more common in the axilla than in the internal mammary chain.

Recently Al-Jurf and Jochimsen reported on two cases of breast edema as a complication of axillary dis- section for malignant m e l a n ~ m a . ' ~ Breast edema de- veloped in one patient shortly after the dissection, but there was complete resolution after six months. Persis- tent breast lymphedema developed in the second patient seven months after surgery.

In this study, lymphedema of the breast was a fre- quently observed complication. Two groups of patients were identified; those with acute or early onset edema and another with a late onset. Early breast edema, ap- pearing before or during external beam radiotherapy, was found to be related to the disruption of lymphatic flow after axillary dissection. Other potential risk fac- tors such as breast and axillary radiation doses, frac- tionation schedules, use of bolus, breast size, weight, and stage could not be shown to be contributory. The edema gradually subsides during the follow-up period and complete resolution may be expected by three years.

Late breast edema occurring 20 months or more post- irradiation was rare in our series. These three patients

100

TABLE 3. Time of Onset of Breast Edema

Treatment

- Severity Before During After

Mild 8* 1 8 t 2 Moderate 2 0 I Severe 0 0 0

3 _ _ _ ~ _ TOTAL 10 18

Five of eight subsequently progressed to moderate edema. t Six of 18 progressed to moderate edema.

received high radiation doses to the axilla, suggesting that a state of chronic and late lymphatic insufficiency had been induced. Neither has shown an improvement, suggesting that this entity may be irreversible.

REFERENCES

I . Calle R, Pilleron J, Schlienger P, Vilcoq. J . Conservative man- agement of operable breast cancer. Cancer 1978; 42:2045 2053.

2. Peters V. Wedge resection with or without radiation in early breast cancer. In1 J Radial Oncol Biol Phys 1977; 2: 1 I5 I - I 156.

3. Spitalier J, Brandone H, Ayme Y, Amalric R, Santamaria F, Seigle J. Cesiumtherapy of breast cancer. A five-year report of 400 consecutive patients. Int J Radial Oncol Biol Phys 1977; 2:231--235.

4. Haagensen CD, Colley E. Radical mastectomy for mammary carcinoma. Ann Surg 1969; 170:884-888.

5 . Urban J . Surgical management of palpable breast cancer. Can- cer 1980; 46:983-987.

6. Hellman S, Harris J , Levene M. Radiation therapy of early carcinoma of the breast without mastectomy. Cancer 1980; 46:988 - 994.

7. Martinez A, Go6net DR. Irradiation with external beam and interstitial radioactive implant as primary treatment for early breast cancer. Surg Gynecol Obsfet: I52:285-290, 198 I .

8. Montague E, Gutierrez A, Barker JL, Tapley ND, Fletcher GH. Conservation surgery and irradiation for the treatment of favorable breast cancer. Cancer 1979; 43: 1058- 106 I .

9. Pierquin B, Owen R. Maylin C el a / . Radical radiation therapy of breast cancer. Ini J Radial Oncol B i d Phys 1980; 6:17-24.

10. Prosnitz L, Goldenberg I, Packard A el al. Radiation therapy as initial treatment for early stage cancer of the breast without mas- tectomy Cancer 1977; 39:917-923.

1 1. Harris J, Levene M, Svensson G, Hellman S. Analysis of cos- metic results following primary radiation therapy for Stages I and 11 carcinoma of the breast. In1 J Radial Oncol Biol Phys 1979; 51257- 261.

12. Svensson G, Bjarngard B, Larsen R, Levene M. A modified three-field technique for breast treatment. In! J Radial Oncol Biol

13. UICC (Union International Contre le Cancer): T N M classi- ficiation of malignant tumours, 3rd ed. Geneva, Switzerland, Impri- merie G. de Buren S. A. Geneva, 1978.

14. Martinez A, Donaldson SS, Bagshaw MA. Special set-up and treatment techniques for the radiotherapy of pediatric malignancies. In1 J Radial Oncol Biol Phys 1977; 2: 1007- I01 6.

15. Cox DR. Regression models and life tables. J R Siar Soc (Series B ) 1972; 34:187-220.

16. Haagensen CD. Anatomy of the mammary gland. The lym- phatics of the breast. I n : Diseases of the Breast, 2nd ed. Philadelphia: WB Saunders, 1971:28-52.

17 . Hulthorn K, Larsson L-G, Ragnhult I . The lymph drainage from the breast to the axillary and parasternal lymph nodes, studied with the aid of colloidal A u ' ~ ' . Aria Radio1 1955; 43:52-64.

18. Turner-Warwick R. The lymphatics of the breast. Br J Surg

19. Al-Jurf A, Jochimsen P. Lymphedema of the breast: A com- plication of axillary dissection for malignant melanoma. Breast Dis Breosr 1979; 5:30-31.

Phys 1980; 6:689-694.

1959; 46:574-582.