- Email: [email protected]
The essential role of radiation therapy in securing locoregional control of merkel cell carcinoma
Int J Radrorron Oncology Biol. Phys.. Vol. Printed I” the U.S.A. All rights reserved.
19. pp. 583-591 Copyright
0360.3016190 $3.00 + .I0 0 1990 Pergamon Press plc
??Original Contribution
THE ESSENTIAL ROLE OF RADIATION THERAPY IN SECURING LOCOREGIONAL CONTROL OF MERKEL CELL CARCINOMA WILLIAM H. MORRISON, CHARLES D. WENDT, M.D.,*
M.D.,* LESTER J. PETERS, M.D.,* ELVIO G. SILVA, M.D.,+ K. KIAN ANG, M.D., PH.D.* AND HELMUTH GOEPFERT, M.D.#
The University of Texas M. D. Anderson Cancer Center, Houston, TX Between 1966 and 1987, 54 patients with non-disseminated Merkel cell carcinoma (MCC) were treated with curative intent at the University of Texas M. D. Anderson Cancer Center. The primary tumor site was in the head and neck in 38 patients. The majority of the patients (57%) were referred with locoregionally recurrent disease. For the whole group, survival was 30% after 5 years. Patients who presented with nodal involvement had a median survival of 13 months compared with 40 months for node-negative patients (p < .04). Only 4/37 patients treated initially by surgery alone were locoregionally controlled, with a median time to recurrence of 4.9 months. Salvage with radiation therapy was attempted in 18 patients (after additional surgery in 14), but was successful in only four. The predominant failure pattern in this subgroup was distant metastases, occurring as a component of initial recurrence in 12/18 patients. Prior to 1982, the philosophy of initial therapy was to give postoperative irradiation only to patients with large primaries or nodal involvement. Subsequently, postoperative radiotherapy has been recommended routinely, and all five patients treated with this approach remain disease-free. In total, 31 patients (including 10 patients with gross disease) were irradiated at M. D. Anderson; only one developed an in-field locoregional recurrence as an initial site of failure. However, three marginal recurrences occurred. The median dose to the primary tumor, first echelon nodes, and supraclavicular nodes was 60, 51, and 50 Gy, respectively. Our current recommendation for initial treatment is excision of the primary tumor followed by irradiation with generous fields to include the primary tumor site and draining regional lymphatics to doses of 46-50 Gy in 2 Gy fractions. For gross unresected disease, 56-60 Gy is recommended. The role of adjuvant systemic therapy remains to be defined. Merkel cell, Radiation therapy, Skin, Neuroendocrine tumor.
INTRODUCTION
U.T. M. D. Anderson Cancer Center (MDA) have focused on the pathological (12, 20, 2 I), surgical (8), and chemotherapeutic (6) aspects of this disease. The present retrospective review was undertaken to assess the role of radiation therapy in MCC patients based on the MDA
the
Merkel cell carcinoma (MCC), also known as endocrine carcinoma of the skin and neuroendocrine carcinoma of the skin, was first described by Toker in 1972 (25). The cells of this dermal malignancy are known to be of neuroendocrine derivation; the exact progenitor cell has not been clearly determined. In 1978, Tang and Toker (22) postulated that the cell of origin was from the neural crest, most probably the dermal neurotactile cell, described by Merkel in 1875 (14). Electron microscopy of MCC cells characteristically shows cytoplasmic extensions and small, uniform, membrane-bound neurosecretory granules (2 1). MCC is a rare tumor, and experience with its management is therefore limited. Several previous reports from
experience, and to compare of other recent reports.
METHODS
our conclusions
with those
AND MATERIALS
Patients
Through December 1987, a total of 56 patients presenting with non-disseminated MCC were identified from the records of the Departments of Radiation Therapy and
Presented at the 3 1st Annual Meeting of the American Society for Therapeutic Radiology and Oncology, San Francisco, 1-6 October 1989. * Dept. of Radiation Therapy. + Dept. of Pathology. * Dept. of Head and Neck Surgery. Reprint requests to: William H. Morrison, M.D., U.T. M. D. Anderson Cancer Center, 15 I5 Holcombe Boulevard, Houston, TX 77030.
Supported in part by Grants CA 06294 and CA 16612 awarded by the National Cancer Institute, U.S. Department of Health and Human Services. Acknowledgements-The authors acknowledge the secretarial assistance of Ms. Joan Day and the data analysis by Ms. Mary Jane Oswald, B.S. Accepted for publication 29 March 1990.
583
584
I. J. Radiation Oncology 0 Biology 0 Physics
September 1990, Volume 19, Number 3
Pathology at M. D. Anderson. The initial patient, accrued in 1966, and subsequent patients until 1979 were thought to have dermal adult neuroblastoma (12). The term MCC was first used at M. D. Anderson in 1979, one year after Tang and Toker (22) proposed that the Merkel cell was the progenitor cell. A retrospective review of all 22 patients previously diagnosed to have adult neuroblastoma was then done, and all were confirmed to have MCC. Patients referred only for evaluation of their pathologic specimens were excluded. Two patients who presented with localized disease received palliative irradiation only because of their debilitated state; they have been excluded, leaving 54 patients suitable for analysis. Twenty-nine patients presented to MDA with recurrent disease, and 23 patients were referred after their initial biopsy for definitive therapy. Two additional patients who presented with nodal involvement (axillary, 1; parotid, 1) without a known skin primary have been included; both had a history of previous excisions for skin carcinomas, and searches for primary lesions in other sites were non-productive. After referral to MDA, 5 1 patients received their subsequent locoregional treatment at our institution; three patients received surgical therapy at MDA and postoperative radiation at other facilities. These three patients, whose radiation therapy records were available, have been included in this report. Prior to referral, 3 of the 54 patients had received radiation therapy to small fields designed to encompass only the primary tumor. Follow-up information was available on all but one patient; this patient was lost to follow-up after being continuously disease-free for 5 years. The original pathologic specimens have been reviewed by one of us (E.G.S.) to confirm the diagnoses. Electron microscopy was used to provide additional diagnostic information on 11 patients. Staging workups in all cases included a history and physical, chest X ray, CBC, and routine serum chemistries. Treatment Patients were divided into three treatment groups, depending upon whether their initial treatment plan was surgery only, radiation therapy only, or surgery and postoperative radiation. Only three patients received chemotherapy as a component of their initial treatment. The surgical procedure used was dependent on the site(s) of involvement. Neck nodal operations were usually modified neck dissections (3) with preservation of the ster-
nocleidomastoid muscle and jugular vein; parotidectomies were done if the primary site drained into those nodes. The majority of the head and neck patients were treated with unilateral electron beam techniques similar to those devised for aggressive squamous skin carcinomas (7, 23). Custom-made bolus and lead shields were frequently used. Preauricular and parotid nodes were treated predominantly with high energy electrons (range: 12- 18 MeV) mixed with photons in a ratio that was typically about 3: 1; the role of the photons was to spare the skin. There was no uniform method of specifying the radiation dose delivered by the electron beam during the period of this study. Doses were therefore recalculated to conform to current practice: the upper neck and parotid fields were specified at the 90% line, and the primary bed and supraclavicular field were designated as given doses. Photon doses were calculated at a depth appropriate to the target volume. Interstitial implants were used for boosts in two patients and as definitive local therapy in one patient. Chemotherapy regimens varied over the years of the study. There were no standardized protocols for MCC chemotherapy, and decisions concerning the indications for chemotherapy and which drugs to use were individualized. Survival was calculated from the date of the initial diagnostic biopsy to the date of the last follow-up or death using the Berkson and Gage life table method (1). RESULTS There were 42 males and 12 females. Fifty of the patients were Caucasian, 3 were Hispanic, and 1 was Black. The median age was 70, with a range from 35 to 9 1. Excluding the two patients with unknown primaries, there were 38 patients whose primary tumor site was in the head and neck; the remaining 14 patients had initial derma1 disease in the trunk (5 patients) or extremities (9 patients). Primary tumors were arbitrarily divided by size into three groups: O-2 cm (42 patients), 2.1-5 cm (7 patients), and larger than 5 cm (3 patients). Nine patients presented with regional nodal involvement prior to any therapy. Table 1 shows the distribution of primary tumor size and presenting nodal status by treatment group. Median freedom from relapse (FFR) and median survival for the whole group were 8 and 38 months, respectively. Overall, FFR and survival were 19% and 30% at 5
Table 1. Treatment
groups Primary
size (cm)
No. pts. N+
Treatment group Surgery alone Surgery + postop Radiation alone
* One additional
XRT
patient
No. pts.
at diagnosis
O-2 cm
2.1-5 cm
>5 cm
37 12 5 54
3 5 1 9
31* 6* 5 42
5 2 0 7
0 3 0 3
in this treatment
group had an unknown
primary
MCC.
Merkel cell carcinoma 0 W.
years (Fig. 1); patients treated initially at MDA had a FFR and survival of 42% and 44%, respectively, at 5 years.
Anal_vsis by primary treatment plan Surgery. Surgical resection (with chemotherapy
in 2 patients) was the treatment for the primary tumor in 36 patients. The initial procedure in these patients was a local excision of the primary with or without a re-excision; the three patients who presented with regional nodal involvement also had a therapeutic nodal dissection (including superficial parotidectomies in 2 patients). Twenty-eight of the 37 patients underwent their initial planned surgical therapy elsewhere and were referred to MDA after a local and/or regional recurrence had occurred. The remaining nine patients were referred to MDA for management immediately after their initial biopsy. The FFR and survival for these 37 patients were 11% and 28%, respectively, at 5 years (Fig. 2). The patterns of recurrence are shown in Table 2. The median time to recurrence was 4.9 months; 9 1% of the recurrences occurred within the first year. Only four of these 37 patients have remained relapse-free, with follow-up intervals of 6, 7, 8, and 8 years. The size of the primary tumors in these four patients was smaller than 2 cm. Thirty-three patients had local and/or regional recurrences after their initial surgical procedures, and 28 of them underwent salvage therapy. The other patients were either too debilitated for salvage therapy or had rapid dissemination of their disease. Eighteen of the 28 patients in whom salvage therapy was attempted received radiation as a component of their treatment. In 14 of these 18 patients, the recurrent disease was excised prior to irradiation. However, four of these patients developed another locoregional recurrence in the interval between surgery and the initiation of radiation. The outcome of salvage therapy as a function of the number of previous recurrences is shown in Table 3. Only four patients survived disease-free; the other 14 patients had the following relapse pattern: recurrence at the radiation field margin (MR) (2 patients), distant metastases (DM) alone (10 patients), DM and MR (1 patient), and DM with probable reseeding of supraclavicular (in-field) nodes (1 patient). The median
__
Overall Survival
-------
Relapse Ree Survival
2
I
I
1
4
6
8
Time in Years
Fig. 1. Freedom from relapse (FFR) and overall survival for all patients.
H. MORRISON et al.
585
survival after radiation therapy for the 18 patients was 12 months. Surgery alone was used in five patients as a salvage procedure; a sixth patient was treated with surgery and chemotherapy. Three of these patients survived long-term (6, 8, and 1 1 years), including the patient treated with chemotherapy. Four patients were treated with chemotherapy alone as a salvage approach; long-term disease control was not achieved in any of these patients. Irradiation. Five patients were treated with radiation alone, and all five relapsed (Table 2). In four patients (who were irradiated prior to 1978), treatment was given to the primary lesion only. Three of these four received their treatment before referral to MDA. One patient (who was treated elsewhere with radium needles for a lip primary in 1966) had a local recurrence. All four patients developed nodal recurrences in unirradiated regional lymphatics. The fifth patient, who received comprehensive radiotherapy, relapsed with distant metastases without locoregional recurrence. Salvage therapies, which included surgery (2 patients) and reirradiation (2 patients), was successful in only one patient. Surger.v and postoperative irradiation. Twelve patients were treated with planned locoregional postoperative radiation after their initial surgery. The first seven patients, treated prior to 1982, underwent excision of the primary tumor and a regional node dissection; postoperative radiotherapy was given for adverse features found in the surgical pathology specimens. These features included pathologically positive nodes (5 patients), an 8 cm primary tumor with a positive surgical margin (1 patient), and a clinically positive neck with an inadequate nodal dissection (1 patient). Two patients with pathologically documented nodal disease (2 and 3 positive nodes) had no further recurrences with follow-up intervals of 5 and 8 years. Distant metastases were the first site of recurrence in two patients. Two patients recurred in unirradiated locoregional sites: these failures were in an unirradiated primary bed surgical scar and in a contralateral unirradiated neck. The final patient, irradiated at another institution, had a local recurrence after treatment with splitcourse radiation (60 Gy in 10 weeks). By 1982, as the natural history of MCC became more clearly understood, postoperative radiation was given electively to 5 of 6 patients referred to MDA after their initial surgical resection. These five patients had primary lesions ranging from 0.5 to 2.2 cm and no clinical evidence of nodal involvement. Therapy included excision of the primary tumor and locoregional irradiation (Fig. 3). Electron beam doses to the primary tumor bed ranged from 50 to 55 Gy in 4-5 weeks; the median nodal dose was 46 Gy given dose (range 45-55 Gy) with a median fraction size of 2.2 Gy (range 2.0-2.5 Gy). These five patients have had no recurrences, with a minimum followup of 3 years. FFR was 56% at 5 years for the postoperative radiation group, compared with 11% for the surgical group (p = ,002) (Fig. 2). Survival at 5 years was 55% for the post-
586
I. J. Radiation Oncology 0 Biology 0 Physics
Surgery -------. Postoperative XRT Radiation
Time
September 1990, Volume 19, Number 3
__ 12
-------
5
b ‘0
in Years
2
# Pts. 37
Surgery Postoperative XRT
&&&ion
4
operative radiation patients versus 28% for the patients treated with surgery. However, the two treatment groups were not comparable; all the postoperative radiation patients received their initial treatment at MDA, whereas 76% of the surgical patients had recurrent disease when they were referred to our institution. On the other hand, a higher percentage of the postoperative radiation patients had initial nodal involvement or large primary tumors (Table 1).
Time in Years
An analysis was made of the radiation doses given to the primary tumor site, upper neck nodes, and lower neck and supraclavicular nodes. Extremity and truncal nodal fields were categorized along with upper neck nodes for analysis in this study. As stated previously, electron doses were specified at the 90% line for the preauricular, parotid, and upper neck nodes and as given doses for primary tumor beds and supraclavicular fields. The doses, fraction sizes, and times used for treatment of the various fields are shown in Table 4. Of the 35 irradiated patients, 31 received radiation treatments at MDA. Only 1 of these 3 1 patients developed an initial recurrence in a radiation field. This relapse occurred after 50 Gy given dose (22 fractions over 40 days using 7 MeV electrons) to a field that had previously been irradiated with two small orthovoltage fields; the patient’s neck had been dissected twice for nodal recurrences. Two patients treated elsewhere had local in-field recurrences: one patient was treated postoperatively with split-course irradiation (60 Gy in 10 weeks) for a vulvar MCC (4) and the other patient had a lip primary treated in 1966 with radium needles (dosimetry not available). There were three recurrences at the margin of the radiation field. An example of a marginal recurrence is shown in Figure 4.
Table 2. First relapse pattern
Surgery alone Surgery + postop Radiation alone
XRT
No. pts. 37 12 5 54
* Irradiated outside MDA (1 patient). + Includes 1 MDA patient with a recurrence * Relapses in unirradiated nodal beds.
grouped
No. relapsed 33 5 5 43
in unirradiated
who
Radiation dose response
Nine patients (Table 1) presented with nodal involvement (N+). Five patients were treated with nodal dissection and postoperative radiation, and a sixth patient received radiation alone. The median survival of the NS patients was 13 months, compared with 40 months for the patients who did not have initial nodal involvement (N-) (p = .04). However, the long-term survival of both the N+ and N- patients was only 22% (Fig. 5). The two N+ patients who remained disease-free had two and three positive neck nodes on presentation and were both treated with postoperative radiation. A survival analysis based on the initial size of the primary was made. As shown in Table 1, larger primary tumors were more often treated with combined modality therapy. Survival at 5 years for those with 12 cm tumors was 1 l/42 patients, compared with 2/7 for the 2.1-5 cm tumor group (p = .8). The three patients with >5 cm tumors had all died within 2 years.
group
I
8
Other parameters analyzed which showed no statistically significant survival differences included primary tumor site (head and neck vs. other sites), sex, and age.
Assessment of clinical prognostic factors
Treatment
5
6
Fig. 2. FFR and survival of patients grouped by their initial therapy. The surgical group includes 28 patients were referred to MDA with locoregional recurrent disease. a: Percent relapse free; b: percent survival.
12
locoregional
by initial treatment Local
Nodal
L and N
9 2*+ 1* 12
17 ::
7 0 0 7
sites.
21
Distant 0 2 1 3
Merkel cell carcinoma 0 W. H. MORRISONet al. Table 3. Results of salvage therapy using radiotherapy +/- excision according to number of previous nest-surgical recurrences
No. previous recurrences
1 1 1 2 3
Site(s) of recurrence
In-field control
No. NED (survival after salvage therapy)
Local Nodal Local and nodal Nodal Local
3*/4 616 515
2 (0.6+, 6 yr) 1 (3+ yr)
l/2 l/l 16/18
1 (8 v)
0 0 4/18
* One supraclavicular failure in setting of widespread inated disease. + Death from intercurrent disease.
dissem-
had gross, macroscopic disease treated at Ten patients MDA with radiation alone. Six patients had dermal tumors (3 of the 6 in the surgical scar), and another three patients had nodal involvement at the time of radiation. One patient had both dermal and nodal disease. All of
587
the dermal lesions treated were smaller than 2 cm. The gross tumor was controlled in the radiation field in these 10 patients using radiation alone. One patient who had multiple 1 cm dermal nodules recurred at the field margin. The doses given to the nodal disease were 60, 60, 65, and 68 Gy. The dermal lesions were treated with 50, 60, 65, 67, 70, 72 (the parotid gland was also treated in this electron field), and 75 Gy (including an interstitial boost). Cause of death Of 54 patients, one patient died from uncontrolled MCC at the primary site. Distant metastases were the cause of death for 32 other patients. There were eight deaths secondary to intercurrent disease. Complications There were no major complications attributable to radiation. Patients experienced the expected acute mucosal reactions which all resolved within a month following treatment. The frequent use of electron beam radiation decreased the volume of normal tissue that received high doses.
Fig. 3. This 72-year-old male presented with a 5 mm MCC lateral to the left eye in 1984. An excisional biopsy was done. The patient was treated with wide-field irradiation to the primary tumor and the ipsilateral draining lymphatics. A: To achieve an adequate margin, the lateral eyelids were treated; the lens was protected with a gold eye shield. The primary field received 52.8 Gy given dose over 5 weeks with 7 MeV electrons. The left parotid was treated with 50 Gy given dose in 25 fractions using 12 MeV electrons, and the left neck received 50 Gy given dose in 5 weeks using 7 MeV electrons. B: Lead cut-outs were used to define the field margins. The patient continues to be disease-free in August 1989.
I. J. Radiation Oncology 0 Biology 0 Physics
September 1990, Volume 19, Number 3
Fig. 4. This 55-year-old male presented in January 1978 with advanced chronic lymphocytic leukemia, scattered squamous cell carcinomas on his face, and a 5 cm MCC of the right cheek. The initial plan was to use orthovoltage to treat the whole face in 2 lateral fields with palliative intent; 5 Gy was given in 5 fractions. The plan was then changed, and the goal became to irradiate the MCC gross disease only. Using 15 MeV electrons, 40 Gy was given in 14 fractions, using tight margins as shown in A. The tumor completely responded within 1 month. B, a photograph taken in May, 1978, shows an extensive recurrence along the anterior and inferior field margin. Because the patient was treated palliatively. he has been excluded from the analysis. The patient died as a result of uncontrolled local disease in June. 1978.
DISCUSSION In our experience, MCC occurs predominantly in elderly white males; however, other large series ( 10, 19) report approximately equal male-to-female ratios. The median age was 70 in this review, and 76% of the primary
-
No Nodal Involvement
#Pt.% 45
%O-
ia
.3
oo-
&40 be
PO-
O!
0
I
2
4
6
Time in Years
Fig. 5. Survival by nodal status at presentation.
8
tumors were located in the head and neck. This represents a substantially higher frequency of head and neck primary tumors than the 48% reported in the literature review by Sibley et al. (19). Table 5 shows local and regional recurrence rates in four series of patients treated primarily with surgery. Local recurrence rates ranged from 30-43%, and nodal relapses occurred in more than half the cases. The average time to develop local recurrence was 4.3 months in the series reported by Meland and Jackson (13); in the series of Bourne et al. (2), the mean interval to regional recurrence was 4 months. At MDA, the median interval to locoregional recurrence was 4.9 months. Our surgical locoregional results are consistent with previous reports. The surgically treated group included 28 patients who were transferred to MDA with recurrent disease; unknown is the number of patients who were treated with surgery, cured, and hence not referred. A bias against surgical locoregional control may have been introduced by including patients referred with recurrent disease. Nevertheless, the results obtained using surgery alone in the different series
Merkel cell carcinoma 0 W. H. MORRISON et al.
589
used to treat: (a) the primary tumor, (b) the upper neck, parotid, and truncal/extremity nodes and (c) the supraclavicular nodes
Table 4. Dose-time-fractionation
Table 4 (a). Given dose to primary
No. pts. 1
I 5 6 6 4 4 27
Median dose (GY)
Fraction size range (GY)
<45 45-49
34 45
2.6 2.8
50-54 55-59 60-64 65-69 70-75*
52 55 60 66 71
2.0-2.5 2.0-2.2 1.8-2.6 2.0-2.2 2.0-2.5
Dose range (GY)
* Includes 2 patients who received boosts with interstitial external beam component of their treatment.
Table 4 (b). Tumor
No. pts. 3 8 7 4 8
1
tumor Median fraction size (GY) 2.1 2.1 2.0 2.1 2.1
implants.
Fraction
Median dose (GY)
<45 45-49 50-54 55-59 60-64 65-69
40.5 45 51 56 60 68
Median time (Days)
19 24
-
28-37 33-39 37-68 38-46 28-43
sizes and times for these patients
dose to upper neck, parotid, and truncal/extremity for electron beams and as tumor dose for photon
Dose range (GY)
Time range (Days)
Fraction size range (GY) 2.0-2.5 1.8-2.3 1.8-2.2
1.8-2.3 2.0-2.1 2.4
nodes expressed fields
36 37 42 44 41
refer only to the
at 90%
Median size (GY)
Time range (Days)
2.0 2.0 2.0 2.0 2.0 -
19-29 29-43 36-43 37-56 39-46 39
Median time (Days) 28 33 37 39 43 -
31
Table 4 (c). Given dose to supraclavicular
Fraction size range (GY)
Median fraction size (GY)
Time range (Days)
Median time (Days)
50 56
2.2-2.25 2.0-2.5 2.0
2.0 -
29-68 27-40 40
36 -
60 65
2.0-2.6 1.97
2.0 -
36-43 46
43 -
No. pts.
Dose range (GY)
Median dose (GY)
2 17 1
45-49 50-54 55-59
3
60-64 65-69
1
region
24
have been disappointing; clearly, there is a role for another modality to improve the locoregional control rate. The radiosensitivity of MCC is well documented in several literature reports. Pacella et al. ( 17) reported that local control of MCC was achieved with radiation in 35 of 36 fields treated in 19 patients; there was measurable disease in 23 of these fields. The median follow-up after radiation therapy in their initial report was only 9 months, but a subsequent update (16) extended the relapse-free observation period by 6 months. In the series of Cotlar et
al. (5), in-field failure occurred in 1 of 8 patients irradiated; the three patients treated with gross disease were locally controlled. O’Brien et al. ( 15) treated seven patients with radiation, and no recurrences occurred within the treatment field. Pilotti et al. (18) irradiated five patients with localized MCC (after surgical excision in 2 patients), and in all five, the disease was controlled. In our series, 31 patients were irradiated at MDA; only one of these patients had an initial in-field local recurrence. Based on these data, it is clear that radiation therapy should play an in-
590
I. J. Radiation Oncology 0 Biology 0 Physics Table 5. Local and regional recurrence rates in patients treated primarily by surgery Author
Sibley et al. (Minnesota) ( 19) Meland and Jackson (Mayo) ( 13) Bourne & O’Rourke (Queensland) (2) Hitchcock et al. (review) ( 10) Present series: surgery only
LR
RR
14/43
(30%)
28/43
(65%)
lo/23
(43%)
12/23
(52%)
10/13
(77%)
8/13
(62%)
80/220 (36%)* 16/37
(43%)
107/193 (55%)+ 24137 (65%)
* Excludes patients with <6 months follow-up if disease is locally controlled. + Excludes patients with < 12 months follow-up if disease is regionally controlled.
tegral role in the locoregional therapy for MCC, as local control can be achieved with irradiation in a high percentage of cases. There were three recurrences at the margin of the radiation field in our series. Knox and Kapp ( 11) and Taxy et al. (24) each reported one marginal recurrence. Other series have not specifically commented on this problem; however, our review of the case histories included in these reports suggests that marginal recurrences did indeed occur (2, 5, 17). It is therefore essential that radiation fields cover the primary tumor, operative bed, and draining lymphatics with generous margins. Over the period of this study, radiation doses used at MDA for MCC were only slightly lower than doses used to treat squamous carcinoma. The in-field local recurrence rate has been very low with the doses used. Since MCC is radiosensitive, it is possible that effective local control could be achieved with significantly lower doses; however, this cannot be proven from our data. Pacella et al. (17) used doses of 45-5 1 Gy in 2.0-2.5 Gy fractions in seven
September 1990, Volume 19, Number 3
patients and reported no local recurrences. They recommended using 50 Gy in 25 fractions over 5 weeks for measurable disease, and proposed that lower doses might be sufficient to control subclinical disease. Currently, we use doses in the range of 56-60 Gy for gross disease (depending on tumor volume) and 46-50 Gy for elective or postoperative treatment of subclinical disease, all in 2 Gy fractions. Using the techniques described, the morbidity of therapy with these doses is minimal; hence, we are reluctant to risk a higher probability of local recurrence by using lower doses. To summarize our treatment strategy, for patients presenting with localized disease, we recommend surgical excision of the primary skin tumor as initial therapy. The excision does not need to be wide, since postoperative radiation is always employed. Radiation therapy using generous fields to cover the primary tumor site, the entire surgical bed, the surgical scar, and draining lymphatics should be initiated without undue delay. For head and neck patients, the draining lymphatics of the whole ipsilateral neck (or bilateral neck for midline lesions) and not just the first echelon nodes should be irradiated. Elective neck dissection is not indicated. For patients who present with regional adenopathy, our practice has been to perform a therapeutic neck dissection and follow this with postoperative irradiation. However, given the radiosensitivity of the tumor, primary radiotherapy to the regional nodes, with dissection reserved for persistent disease, would be an appropriate alternative approach. The results achieved with chemotherapy for distant metastatic disease in our institution have recently been published (6). Of 13 patients treated, 3 patients had a complete response to therapy. One patient, whose biopsyproven axillary disease was treated with chemotherapy alone, has been disease-free for 8 years. A few chemotherapy series (9, 26) report modestly encouraging results in small numbers of patients with metastatic disease. The ability of chemotherapy administered in the adjuvant setting to control subclinical MCC remains to be defined.
REFERENCES Berkson, J.; Gage, R. Calculation of survival rates for cancer. Proc. Mayo Clinic 25:270; 1950. Bourne, R. G.; O’Rourke, M. G. E. Management of Merkel cell tumour. Aust. N.Z. J. Surg. 58:97 l-974; 1988. Byers, R. M. Modified neck dissection: a study of 967 cases from 1970 to 1980. Am. J. Surg. 150:414-421; 1985. Copeland, L. J.; Cleary, K.; Sneige, N.; Edwards, C. L. Neuroendocrine (Merkel cell) carcinoma of the vulva: a case report and review of the literature. Gynecol. Oncol. 22:367378; 1985. 5. Cotlar, A. M.; Gates, J. 0.; Gibbs, F. A., Jr. Merkel cell carcinoma: combined surgery and radiation therapy. Am. Surg. 52:159-164; 1986. 6. Feun, L. G.; Savaraj, N.; Legha, S. S.; Silva, E.G.; Benjamin, R. S.; Burgess, M. A. Chemotherapy for metastatic Merkel cell carcinoma: review of the M. D. Anderson Hospital’s experience. Cancer 62:683-685; 1988.
7. Fletcher, G. H. Textbook of radiotherapy, 3rd ed. Philadelphia: Lea & Febiger; 1980. 8. Goepfert, H.; Remmler, D.; Silva, E.; Wheeler, B. Merkel cell carcinoma (endocrine carcinoma of the skin) of the head and neck. Arch. Otolaryngol. 110:707-7 12; 1984. 9. Grosh, W. W.; Giannone, L.; Hande, K. R.; Johnson, D. H. Disseminated Merkel cell tumor: treatment with systemic chemotherapy. Am. J. Clin. Oncol. 10:227-230; 1987. 10. Hitchcock, C. L.; Bland, K. I.; Laney, R. G., III; Franzini, D.; Harris, B.; Copeland, E. M., III. Neuroendocrine (Merkel cell) carcinoma of the skin. Its natural history, diagnosis, and treatment. Ann. Surg. 207:201-207; 1988. 11. Knox, S. J.; Kapp, D. S. Hyperthermia and radiation therapy in the treatment of recurrent Merkel cell tumors. Cancer 62:1479-1486; 1988. 12. Mackay, B.; Luna, M. A.: Butler, J. J. Adult neuroblastoma:
Merkel cell carcinoma 0 W. H. MORRISONet al.
13.
14.
15.
16.
17.
18.
electron microscopic observations in nine cases. Cancer 37: 1334-1351: 1976. Meland, N. B.; Jackson, 1. T. Merkel cell tumor: diagnosis, prognosis, and management. Plast. Reconstr. Surg. 77:632638: 1986. Merkel, F. Tastzellen und Tastkoerperchen bei den haushieren und beim menschen. Arch. Mikr. Anat. 11:636-652; 1875. O’Brien, P. C.; Denham, J. W.; Leong, S-Y. Merkel cell carcinoma: a review of behaviour patterns and management strategies. Aust. N. Z. J. Surg. 57:847-850: 1987. Pacella, J. A.; Ashby, M. A. Response to editorial “Neuroendocrine carcinoma of the skin: diagnostic and management considerations” (Letter). Int. J. Radiat. Oncol. Biol. Phys. 15:249: 1988. Pacella, J. A.; Ashby, M. A.: Ainslie, J.; Minty, C. The role of radiotherapy in the management of primary cutaneous neuroendocrine tumors (Merkel cell or trabecular carcinoma): experience at the Peter MacCallum Cancer Institute (Melbourne, Australia). Int. J. Radiat. Oncol. Biol. Phys. 14:1077-1084; 1988. Pilotti, S.; Rilke. F.; Bartoli, C.: Grisotti, A. Clinicopathologic
19.
20.
2 1.
22. 23. 24. 25. 26.
591
correlations of cutaneous neuroendocrine Merkel cell carcinoma. J. Clin. Oncol. 6: 1863-1873; 1988. Sibley, R. K.; Dehner, L. P.; Rosai, J. Primary neuroendocrine (Merkel cell?) carcinoma of the skin. I. A clinicopathologic and ultrastructural study of 43 cases. Am. J. Surg. Pathol. 9:95-108; 1985. Silva, E.; Mackay, B. Neuroendocrine (Merkel cell) carcinomas of the skin: an ultrastructural study of nine cases. Ultrastruct. Pathol. 2: l-9; 198 1. Silva, E. G.; Mackay, B.; Goepfert, H.; Burgess, M. A.; Fields, R. S. Endocrine carcinoma of the skin (Merkel cell carcinoma). Pathol. Annu. l9:1-30; 1984. Tang, C-K.: Toker, C. Trabecular carcinoma of the skin. An ultrastructural study. Cancer 42:23 I l-232 I ; 1978. Tapley, N. duV. Clinical applications ofthe electron beam. New York: John Wiley & sons; 1976. Taxy, J. B.; Ettinger, D. S.; Wharam, M. D. Primary small cell carcinoma of the skin. Cancer 46:2308-231 I: 1980. Toker. C. Trabecular carcinoma of the skin. Arch. Dermatol. 105:107-l IO; 1972. Wynne, C. J.: Kearsley, J. H. Merkel cell tumor: a chemosensitive skin cancer. Cancer 62:28-3 1; 1988.
19. pp. 583-591 Copyright
0360.3016190 $3.00 + .I0 0 1990 Pergamon Press plc
??Original Contribution
THE ESSENTIAL ROLE OF RADIATION THERAPY IN SECURING LOCOREGIONAL CONTROL OF MERKEL CELL CARCINOMA WILLIAM H. MORRISON, CHARLES D. WENDT, M.D.,*
M.D.,* LESTER J. PETERS, M.D.,* ELVIO G. SILVA, M.D.,+ K. KIAN ANG, M.D., PH.D.* AND HELMUTH GOEPFERT, M.D.#
The University of Texas M. D. Anderson Cancer Center, Houston, TX Between 1966 and 1987, 54 patients with non-disseminated Merkel cell carcinoma (MCC) were treated with curative intent at the University of Texas M. D. Anderson Cancer Center. The primary tumor site was in the head and neck in 38 patients. The majority of the patients (57%) were referred with locoregionally recurrent disease. For the whole group, survival was 30% after 5 years. Patients who presented with nodal involvement had a median survival of 13 months compared with 40 months for node-negative patients (p < .04). Only 4/37 patients treated initially by surgery alone were locoregionally controlled, with a median time to recurrence of 4.9 months. Salvage with radiation therapy was attempted in 18 patients (after additional surgery in 14), but was successful in only four. The predominant failure pattern in this subgroup was distant metastases, occurring as a component of initial recurrence in 12/18 patients. Prior to 1982, the philosophy of initial therapy was to give postoperative irradiation only to patients with large primaries or nodal involvement. Subsequently, postoperative radiotherapy has been recommended routinely, and all five patients treated with this approach remain disease-free. In total, 31 patients (including 10 patients with gross disease) were irradiated at M. D. Anderson; only one developed an in-field locoregional recurrence as an initial site of failure. However, three marginal recurrences occurred. The median dose to the primary tumor, first echelon nodes, and supraclavicular nodes was 60, 51, and 50 Gy, respectively. Our current recommendation for initial treatment is excision of the primary tumor followed by irradiation with generous fields to include the primary tumor site and draining regional lymphatics to doses of 46-50 Gy in 2 Gy fractions. For gross unresected disease, 56-60 Gy is recommended. The role of adjuvant systemic therapy remains to be defined. Merkel cell, Radiation therapy, Skin, Neuroendocrine tumor.
INTRODUCTION
U.T. M. D. Anderson Cancer Center (MDA) have focused on the pathological (12, 20, 2 I), surgical (8), and chemotherapeutic (6) aspects of this disease. The present retrospective review was undertaken to assess the role of radiation therapy in MCC patients based on the MDA
the
Merkel cell carcinoma (MCC), also known as endocrine carcinoma of the skin and neuroendocrine carcinoma of the skin, was first described by Toker in 1972 (25). The cells of this dermal malignancy are known to be of neuroendocrine derivation; the exact progenitor cell has not been clearly determined. In 1978, Tang and Toker (22) postulated that the cell of origin was from the neural crest, most probably the dermal neurotactile cell, described by Merkel in 1875 (14). Electron microscopy of MCC cells characteristically shows cytoplasmic extensions and small, uniform, membrane-bound neurosecretory granules (2 1). MCC is a rare tumor, and experience with its management is therefore limited. Several previous reports from
experience, and to compare of other recent reports.
METHODS
our conclusions
with those
AND MATERIALS
Patients
Through December 1987, a total of 56 patients presenting with non-disseminated MCC were identified from the records of the Departments of Radiation Therapy and
Presented at the 3 1st Annual Meeting of the American Society for Therapeutic Radiology and Oncology, San Francisco, 1-6 October 1989. * Dept. of Radiation Therapy. + Dept. of Pathology. * Dept. of Head and Neck Surgery. Reprint requests to: William H. Morrison, M.D., U.T. M. D. Anderson Cancer Center, 15 I5 Holcombe Boulevard, Houston, TX 77030.
Supported in part by Grants CA 06294 and CA 16612 awarded by the National Cancer Institute, U.S. Department of Health and Human Services. Acknowledgements-The authors acknowledge the secretarial assistance of Ms. Joan Day and the data analysis by Ms. Mary Jane Oswald, B.S. Accepted for publication 29 March 1990.
583
584
I. J. Radiation Oncology 0 Biology 0 Physics
September 1990, Volume 19, Number 3
Pathology at M. D. Anderson. The initial patient, accrued in 1966, and subsequent patients until 1979 were thought to have dermal adult neuroblastoma (12). The term MCC was first used at M. D. Anderson in 1979, one year after Tang and Toker (22) proposed that the Merkel cell was the progenitor cell. A retrospective review of all 22 patients previously diagnosed to have adult neuroblastoma was then done, and all were confirmed to have MCC. Patients referred only for evaluation of their pathologic specimens were excluded. Two patients who presented with localized disease received palliative irradiation only because of their debilitated state; they have been excluded, leaving 54 patients suitable for analysis. Twenty-nine patients presented to MDA with recurrent disease, and 23 patients were referred after their initial biopsy for definitive therapy. Two additional patients who presented with nodal involvement (axillary, 1; parotid, 1) without a known skin primary have been included; both had a history of previous excisions for skin carcinomas, and searches for primary lesions in other sites were non-productive. After referral to MDA, 5 1 patients received their subsequent locoregional treatment at our institution; three patients received surgical therapy at MDA and postoperative radiation at other facilities. These three patients, whose radiation therapy records were available, have been included in this report. Prior to referral, 3 of the 54 patients had received radiation therapy to small fields designed to encompass only the primary tumor. Follow-up information was available on all but one patient; this patient was lost to follow-up after being continuously disease-free for 5 years. The original pathologic specimens have been reviewed by one of us (E.G.S.) to confirm the diagnoses. Electron microscopy was used to provide additional diagnostic information on 11 patients. Staging workups in all cases included a history and physical, chest X ray, CBC, and routine serum chemistries. Treatment Patients were divided into three treatment groups, depending upon whether their initial treatment plan was surgery only, radiation therapy only, or surgery and postoperative radiation. Only three patients received chemotherapy as a component of their initial treatment. The surgical procedure used was dependent on the site(s) of involvement. Neck nodal operations were usually modified neck dissections (3) with preservation of the ster-
nocleidomastoid muscle and jugular vein; parotidectomies were done if the primary site drained into those nodes. The majority of the head and neck patients were treated with unilateral electron beam techniques similar to those devised for aggressive squamous skin carcinomas (7, 23). Custom-made bolus and lead shields were frequently used. Preauricular and parotid nodes were treated predominantly with high energy electrons (range: 12- 18 MeV) mixed with photons in a ratio that was typically about 3: 1; the role of the photons was to spare the skin. There was no uniform method of specifying the radiation dose delivered by the electron beam during the period of this study. Doses were therefore recalculated to conform to current practice: the upper neck and parotid fields were specified at the 90% line, and the primary bed and supraclavicular field were designated as given doses. Photon doses were calculated at a depth appropriate to the target volume. Interstitial implants were used for boosts in two patients and as definitive local therapy in one patient. Chemotherapy regimens varied over the years of the study. There were no standardized protocols for MCC chemotherapy, and decisions concerning the indications for chemotherapy and which drugs to use were individualized. Survival was calculated from the date of the initial diagnostic biopsy to the date of the last follow-up or death using the Berkson and Gage life table method (1). RESULTS There were 42 males and 12 females. Fifty of the patients were Caucasian, 3 were Hispanic, and 1 was Black. The median age was 70, with a range from 35 to 9 1. Excluding the two patients with unknown primaries, there were 38 patients whose primary tumor site was in the head and neck; the remaining 14 patients had initial derma1 disease in the trunk (5 patients) or extremities (9 patients). Primary tumors were arbitrarily divided by size into three groups: O-2 cm (42 patients), 2.1-5 cm (7 patients), and larger than 5 cm (3 patients). Nine patients presented with regional nodal involvement prior to any therapy. Table 1 shows the distribution of primary tumor size and presenting nodal status by treatment group. Median freedom from relapse (FFR) and median survival for the whole group were 8 and 38 months, respectively. Overall, FFR and survival were 19% and 30% at 5
Table 1. Treatment
groups Primary
size (cm)
No. pts. N+
Treatment group Surgery alone Surgery + postop Radiation alone
* One additional
XRT
patient
No. pts.
at diagnosis
O-2 cm
2.1-5 cm
>5 cm
37 12 5 54
3 5 1 9
31* 6* 5 42
5 2 0 7
0 3 0 3
in this treatment
group had an unknown
primary
MCC.
Merkel cell carcinoma 0 W.
years (Fig. 1); patients treated initially at MDA had a FFR and survival of 42% and 44%, respectively, at 5 years.
Anal_vsis by primary treatment plan Surgery. Surgical resection (with chemotherapy
in 2 patients) was the treatment for the primary tumor in 36 patients. The initial procedure in these patients was a local excision of the primary with or without a re-excision; the three patients who presented with regional nodal involvement also had a therapeutic nodal dissection (including superficial parotidectomies in 2 patients). Twenty-eight of the 37 patients underwent their initial planned surgical therapy elsewhere and were referred to MDA after a local and/or regional recurrence had occurred. The remaining nine patients were referred to MDA for management immediately after their initial biopsy. The FFR and survival for these 37 patients were 11% and 28%, respectively, at 5 years (Fig. 2). The patterns of recurrence are shown in Table 2. The median time to recurrence was 4.9 months; 9 1% of the recurrences occurred within the first year. Only four of these 37 patients have remained relapse-free, with follow-up intervals of 6, 7, 8, and 8 years. The size of the primary tumors in these four patients was smaller than 2 cm. Thirty-three patients had local and/or regional recurrences after their initial surgical procedures, and 28 of them underwent salvage therapy. The other patients were either too debilitated for salvage therapy or had rapid dissemination of their disease. Eighteen of the 28 patients in whom salvage therapy was attempted received radiation as a component of their treatment. In 14 of these 18 patients, the recurrent disease was excised prior to irradiation. However, four of these patients developed another locoregional recurrence in the interval between surgery and the initiation of radiation. The outcome of salvage therapy as a function of the number of previous recurrences is shown in Table 3. Only four patients survived disease-free; the other 14 patients had the following relapse pattern: recurrence at the radiation field margin (MR) (2 patients), distant metastases (DM) alone (10 patients), DM and MR (1 patient), and DM with probable reseeding of supraclavicular (in-field) nodes (1 patient). The median
__
Overall Survival
-------
Relapse Ree Survival
2
I
I
1
4
6
8
Time in Years
Fig. 1. Freedom from relapse (FFR) and overall survival for all patients.
H. MORRISON et al.
585
survival after radiation therapy for the 18 patients was 12 months. Surgery alone was used in five patients as a salvage procedure; a sixth patient was treated with surgery and chemotherapy. Three of these patients survived long-term (6, 8, and 1 1 years), including the patient treated with chemotherapy. Four patients were treated with chemotherapy alone as a salvage approach; long-term disease control was not achieved in any of these patients. Irradiation. Five patients were treated with radiation alone, and all five relapsed (Table 2). In four patients (who were irradiated prior to 1978), treatment was given to the primary lesion only. Three of these four received their treatment before referral to MDA. One patient (who was treated elsewhere with radium needles for a lip primary in 1966) had a local recurrence. All four patients developed nodal recurrences in unirradiated regional lymphatics. The fifth patient, who received comprehensive radiotherapy, relapsed with distant metastases without locoregional recurrence. Salvage therapies, which included surgery (2 patients) and reirradiation (2 patients), was successful in only one patient. Surger.v and postoperative irradiation. Twelve patients were treated with planned locoregional postoperative radiation after their initial surgery. The first seven patients, treated prior to 1982, underwent excision of the primary tumor and a regional node dissection; postoperative radiotherapy was given for adverse features found in the surgical pathology specimens. These features included pathologically positive nodes (5 patients), an 8 cm primary tumor with a positive surgical margin (1 patient), and a clinically positive neck with an inadequate nodal dissection (1 patient). Two patients with pathologically documented nodal disease (2 and 3 positive nodes) had no further recurrences with follow-up intervals of 5 and 8 years. Distant metastases were the first site of recurrence in two patients. Two patients recurred in unirradiated locoregional sites: these failures were in an unirradiated primary bed surgical scar and in a contralateral unirradiated neck. The final patient, irradiated at another institution, had a local recurrence after treatment with splitcourse radiation (60 Gy in 10 weeks). By 1982, as the natural history of MCC became more clearly understood, postoperative radiation was given electively to 5 of 6 patients referred to MDA after their initial surgical resection. These five patients had primary lesions ranging from 0.5 to 2.2 cm and no clinical evidence of nodal involvement. Therapy included excision of the primary tumor and locoregional irradiation (Fig. 3). Electron beam doses to the primary tumor bed ranged from 50 to 55 Gy in 4-5 weeks; the median nodal dose was 46 Gy given dose (range 45-55 Gy) with a median fraction size of 2.2 Gy (range 2.0-2.5 Gy). These five patients have had no recurrences, with a minimum followup of 3 years. FFR was 56% at 5 years for the postoperative radiation group, compared with 11% for the surgical group (p = ,002) (Fig. 2). Survival at 5 years was 55% for the post-
586
I. J. Radiation Oncology 0 Biology 0 Physics
Surgery -------. Postoperative XRT Radiation
Time
September 1990, Volume 19, Number 3
__ 12
-------
5
b ‘0
in Years
2
# Pts. 37
Surgery Postoperative XRT
&&&ion
4
operative radiation patients versus 28% for the patients treated with surgery. However, the two treatment groups were not comparable; all the postoperative radiation patients received their initial treatment at MDA, whereas 76% of the surgical patients had recurrent disease when they were referred to our institution. On the other hand, a higher percentage of the postoperative radiation patients had initial nodal involvement or large primary tumors (Table 1).
Time in Years
An analysis was made of the radiation doses given to the primary tumor site, upper neck nodes, and lower neck and supraclavicular nodes. Extremity and truncal nodal fields were categorized along with upper neck nodes for analysis in this study. As stated previously, electron doses were specified at the 90% line for the preauricular, parotid, and upper neck nodes and as given doses for primary tumor beds and supraclavicular fields. The doses, fraction sizes, and times used for treatment of the various fields are shown in Table 4. Of the 35 irradiated patients, 31 received radiation treatments at MDA. Only 1 of these 3 1 patients developed an initial recurrence in a radiation field. This relapse occurred after 50 Gy given dose (22 fractions over 40 days using 7 MeV electrons) to a field that had previously been irradiated with two small orthovoltage fields; the patient’s neck had been dissected twice for nodal recurrences. Two patients treated elsewhere had local in-field recurrences: one patient was treated postoperatively with split-course irradiation (60 Gy in 10 weeks) for a vulvar MCC (4) and the other patient had a lip primary treated in 1966 with radium needles (dosimetry not available). There were three recurrences at the margin of the radiation field. An example of a marginal recurrence is shown in Figure 4.
Table 2. First relapse pattern
Surgery alone Surgery + postop Radiation alone
XRT
No. pts. 37 12 5 54
* Irradiated outside MDA (1 patient). + Includes 1 MDA patient with a recurrence * Relapses in unirradiated nodal beds.
grouped
No. relapsed 33 5 5 43
in unirradiated
who
Radiation dose response
Nine patients (Table 1) presented with nodal involvement (N+). Five patients were treated with nodal dissection and postoperative radiation, and a sixth patient received radiation alone. The median survival of the NS patients was 13 months, compared with 40 months for the patients who did not have initial nodal involvement (N-) (p = .04). However, the long-term survival of both the N+ and N- patients was only 22% (Fig. 5). The two N+ patients who remained disease-free had two and three positive neck nodes on presentation and were both treated with postoperative radiation. A survival analysis based on the initial size of the primary was made. As shown in Table 1, larger primary tumors were more often treated with combined modality therapy. Survival at 5 years for those with 12 cm tumors was 1 l/42 patients, compared with 2/7 for the 2.1-5 cm tumor group (p = .8). The three patients with >5 cm tumors had all died within 2 years.
group
I
8
Other parameters analyzed which showed no statistically significant survival differences included primary tumor site (head and neck vs. other sites), sex, and age.
Assessment of clinical prognostic factors
Treatment
5
6
Fig. 2. FFR and survival of patients grouped by their initial therapy. The surgical group includes 28 patients were referred to MDA with locoregional recurrent disease. a: Percent relapse free; b: percent survival.
12
locoregional
by initial treatment Local
Nodal
L and N
9 2*+ 1* 12
17 ::
7 0 0 7
sites.
21
Distant 0 2 1 3
Merkel cell carcinoma 0 W. H. MORRISONet al. Table 3. Results of salvage therapy using radiotherapy +/- excision according to number of previous nest-surgical recurrences
No. previous recurrences
1 1 1 2 3
Site(s) of recurrence
In-field control
No. NED (survival after salvage therapy)
Local Nodal Local and nodal Nodal Local
3*/4 616 515
2 (0.6+, 6 yr) 1 (3+ yr)
l/2 l/l 16/18
1 (8 v)
0 0 4/18
* One supraclavicular failure in setting of widespread inated disease. + Death from intercurrent disease.
dissem-
had gross, macroscopic disease treated at Ten patients MDA with radiation alone. Six patients had dermal tumors (3 of the 6 in the surgical scar), and another three patients had nodal involvement at the time of radiation. One patient had both dermal and nodal disease. All of
587
the dermal lesions treated were smaller than 2 cm. The gross tumor was controlled in the radiation field in these 10 patients using radiation alone. One patient who had multiple 1 cm dermal nodules recurred at the field margin. The doses given to the nodal disease were 60, 60, 65, and 68 Gy. The dermal lesions were treated with 50, 60, 65, 67, 70, 72 (the parotid gland was also treated in this electron field), and 75 Gy (including an interstitial boost). Cause of death Of 54 patients, one patient died from uncontrolled MCC at the primary site. Distant metastases were the cause of death for 32 other patients. There were eight deaths secondary to intercurrent disease. Complications There were no major complications attributable to radiation. Patients experienced the expected acute mucosal reactions which all resolved within a month following treatment. The frequent use of electron beam radiation decreased the volume of normal tissue that received high doses.
Fig. 3. This 72-year-old male presented with a 5 mm MCC lateral to the left eye in 1984. An excisional biopsy was done. The patient was treated with wide-field irradiation to the primary tumor and the ipsilateral draining lymphatics. A: To achieve an adequate margin, the lateral eyelids were treated; the lens was protected with a gold eye shield. The primary field received 52.8 Gy given dose over 5 weeks with 7 MeV electrons. The left parotid was treated with 50 Gy given dose in 25 fractions using 12 MeV electrons, and the left neck received 50 Gy given dose in 5 weeks using 7 MeV electrons. B: Lead cut-outs were used to define the field margins. The patient continues to be disease-free in August 1989.
I. J. Radiation Oncology 0 Biology 0 Physics
September 1990, Volume 19, Number 3
Fig. 4. This 55-year-old male presented in January 1978 with advanced chronic lymphocytic leukemia, scattered squamous cell carcinomas on his face, and a 5 cm MCC of the right cheek. The initial plan was to use orthovoltage to treat the whole face in 2 lateral fields with palliative intent; 5 Gy was given in 5 fractions. The plan was then changed, and the goal became to irradiate the MCC gross disease only. Using 15 MeV electrons, 40 Gy was given in 14 fractions, using tight margins as shown in A. The tumor completely responded within 1 month. B, a photograph taken in May, 1978, shows an extensive recurrence along the anterior and inferior field margin. Because the patient was treated palliatively. he has been excluded from the analysis. The patient died as a result of uncontrolled local disease in June. 1978.
DISCUSSION In our experience, MCC occurs predominantly in elderly white males; however, other large series ( 10, 19) report approximately equal male-to-female ratios. The median age was 70 in this review, and 76% of the primary
-
No Nodal Involvement
#Pt.% 45
%O-
ia
.3
oo-
&40 be
PO-
O!
0
I
2
4
6
Time in Years
Fig. 5. Survival by nodal status at presentation.
8
tumors were located in the head and neck. This represents a substantially higher frequency of head and neck primary tumors than the 48% reported in the literature review by Sibley et al. (19). Table 5 shows local and regional recurrence rates in four series of patients treated primarily with surgery. Local recurrence rates ranged from 30-43%, and nodal relapses occurred in more than half the cases. The average time to develop local recurrence was 4.3 months in the series reported by Meland and Jackson (13); in the series of Bourne et al. (2), the mean interval to regional recurrence was 4 months. At MDA, the median interval to locoregional recurrence was 4.9 months. Our surgical locoregional results are consistent with previous reports. The surgically treated group included 28 patients who were transferred to MDA with recurrent disease; unknown is the number of patients who were treated with surgery, cured, and hence not referred. A bias against surgical locoregional control may have been introduced by including patients referred with recurrent disease. Nevertheless, the results obtained using surgery alone in the different series
Merkel cell carcinoma 0 W. H. MORRISON et al.
589
used to treat: (a) the primary tumor, (b) the upper neck, parotid, and truncal/extremity nodes and (c) the supraclavicular nodes
Table 4. Dose-time-fractionation
Table 4 (a). Given dose to primary
No. pts. 1
I 5 6 6 4 4 27
Median dose (GY)
Fraction size range (GY)
<45 45-49
34 45
2.6 2.8
50-54 55-59 60-64 65-69 70-75*
52 55 60 66 71
2.0-2.5 2.0-2.2 1.8-2.6 2.0-2.2 2.0-2.5
Dose range (GY)
* Includes 2 patients who received boosts with interstitial external beam component of their treatment.
Table 4 (b). Tumor
No. pts. 3 8 7 4 8
1
tumor Median fraction size (GY) 2.1 2.1 2.0 2.1 2.1
implants.
Fraction
Median dose (GY)
<45 45-49 50-54 55-59 60-64 65-69
40.5 45 51 56 60 68
Median time (Days)
19 24
-
28-37 33-39 37-68 38-46 28-43
sizes and times for these patients
dose to upper neck, parotid, and truncal/extremity for electron beams and as tumor dose for photon
Dose range (GY)
Time range (Days)
Fraction size range (GY) 2.0-2.5 1.8-2.3 1.8-2.2
1.8-2.3 2.0-2.1 2.4
nodes expressed fields
36 37 42 44 41
refer only to the
at 90%
Median size (GY)
Time range (Days)
2.0 2.0 2.0 2.0 2.0 -
19-29 29-43 36-43 37-56 39-46 39
Median time (Days) 28 33 37 39 43 -
31
Table 4 (c). Given dose to supraclavicular
Fraction size range (GY)
Median fraction size (GY)
Time range (Days)
Median time (Days)
50 56
2.2-2.25 2.0-2.5 2.0
2.0 -
29-68 27-40 40
36 -
60 65
2.0-2.6 1.97
2.0 -
36-43 46
43 -
No. pts.
Dose range (GY)
Median dose (GY)
2 17 1
45-49 50-54 55-59
3
60-64 65-69
1
region
24
have been disappointing; clearly, there is a role for another modality to improve the locoregional control rate. The radiosensitivity of MCC is well documented in several literature reports. Pacella et al. ( 17) reported that local control of MCC was achieved with radiation in 35 of 36 fields treated in 19 patients; there was measurable disease in 23 of these fields. The median follow-up after radiation therapy in their initial report was only 9 months, but a subsequent update (16) extended the relapse-free observation period by 6 months. In the series of Cotlar et
al. (5), in-field failure occurred in 1 of 8 patients irradiated; the three patients treated with gross disease were locally controlled. O’Brien et al. ( 15) treated seven patients with radiation, and no recurrences occurred within the treatment field. Pilotti et al. (18) irradiated five patients with localized MCC (after surgical excision in 2 patients), and in all five, the disease was controlled. In our series, 31 patients were irradiated at MDA; only one of these patients had an initial in-field local recurrence. Based on these data, it is clear that radiation therapy should play an in-
590
I. J. Radiation Oncology 0 Biology 0 Physics Table 5. Local and regional recurrence rates in patients treated primarily by surgery Author
Sibley et al. (Minnesota) ( 19) Meland and Jackson (Mayo) ( 13) Bourne & O’Rourke (Queensland) (2) Hitchcock et al. (review) ( 10) Present series: surgery only
LR
RR
14/43
(30%)
28/43
(65%)
lo/23
(43%)
12/23
(52%)
10/13
(77%)
8/13
(62%)
80/220 (36%)* 16/37
(43%)
107/193 (55%)+ 24137 (65%)
* Excludes patients with <6 months follow-up if disease is locally controlled. + Excludes patients with < 12 months follow-up if disease is regionally controlled.
tegral role in the locoregional therapy for MCC, as local control can be achieved with irradiation in a high percentage of cases. There were three recurrences at the margin of the radiation field in our series. Knox and Kapp ( 11) and Taxy et al. (24) each reported one marginal recurrence. Other series have not specifically commented on this problem; however, our review of the case histories included in these reports suggests that marginal recurrences did indeed occur (2, 5, 17). It is therefore essential that radiation fields cover the primary tumor, operative bed, and draining lymphatics with generous margins. Over the period of this study, radiation doses used at MDA for MCC were only slightly lower than doses used to treat squamous carcinoma. The in-field local recurrence rate has been very low with the doses used. Since MCC is radiosensitive, it is possible that effective local control could be achieved with significantly lower doses; however, this cannot be proven from our data. Pacella et al. (17) used doses of 45-5 1 Gy in 2.0-2.5 Gy fractions in seven
September 1990, Volume 19, Number 3
patients and reported no local recurrences. They recommended using 50 Gy in 25 fractions over 5 weeks for measurable disease, and proposed that lower doses might be sufficient to control subclinical disease. Currently, we use doses in the range of 56-60 Gy for gross disease (depending on tumor volume) and 46-50 Gy for elective or postoperative treatment of subclinical disease, all in 2 Gy fractions. Using the techniques described, the morbidity of therapy with these doses is minimal; hence, we are reluctant to risk a higher probability of local recurrence by using lower doses. To summarize our treatment strategy, for patients presenting with localized disease, we recommend surgical excision of the primary skin tumor as initial therapy. The excision does not need to be wide, since postoperative radiation is always employed. Radiation therapy using generous fields to cover the primary tumor site, the entire surgical bed, the surgical scar, and draining lymphatics should be initiated without undue delay. For head and neck patients, the draining lymphatics of the whole ipsilateral neck (or bilateral neck for midline lesions) and not just the first echelon nodes should be irradiated. Elective neck dissection is not indicated. For patients who present with regional adenopathy, our practice has been to perform a therapeutic neck dissection and follow this with postoperative irradiation. However, given the radiosensitivity of the tumor, primary radiotherapy to the regional nodes, with dissection reserved for persistent disease, would be an appropriate alternative approach. The results achieved with chemotherapy for distant metastatic disease in our institution have recently been published (6). Of 13 patients treated, 3 patients had a complete response to therapy. One patient, whose biopsyproven axillary disease was treated with chemotherapy alone, has been disease-free for 8 years. A few chemotherapy series (9, 26) report modestly encouraging results in small numbers of patients with metastatic disease. The ability of chemotherapy administered in the adjuvant setting to control subclinical MCC remains to be defined.
REFERENCES Berkson, J.; Gage, R. Calculation of survival rates for cancer. Proc. Mayo Clinic 25:270; 1950. Bourne, R. G.; O’Rourke, M. G. E. Management of Merkel cell tumour. Aust. N.Z. J. Surg. 58:97 l-974; 1988. Byers, R. M. Modified neck dissection: a study of 967 cases from 1970 to 1980. Am. J. Surg. 150:414-421; 1985. Copeland, L. J.; Cleary, K.; Sneige, N.; Edwards, C. L. Neuroendocrine (Merkel cell) carcinoma of the vulva: a case report and review of the literature. Gynecol. Oncol. 22:367378; 1985. 5. Cotlar, A. M.; Gates, J. 0.; Gibbs, F. A., Jr. Merkel cell carcinoma: combined surgery and radiation therapy. Am. Surg. 52:159-164; 1986. 6. Feun, L. G.; Savaraj, N.; Legha, S. S.; Silva, E.G.; Benjamin, R. S.; Burgess, M. A. Chemotherapy for metastatic Merkel cell carcinoma: review of the M. D. Anderson Hospital’s experience. Cancer 62:683-685; 1988.
7. Fletcher, G. H. Textbook of radiotherapy, 3rd ed. Philadelphia: Lea & Febiger; 1980. 8. Goepfert, H.; Remmler, D.; Silva, E.; Wheeler, B. Merkel cell carcinoma (endocrine carcinoma of the skin) of the head and neck. Arch. Otolaryngol. 110:707-7 12; 1984. 9. Grosh, W. W.; Giannone, L.; Hande, K. R.; Johnson, D. H. Disseminated Merkel cell tumor: treatment with systemic chemotherapy. Am. J. Clin. Oncol. 10:227-230; 1987. 10. Hitchcock, C. L.; Bland, K. I.; Laney, R. G., III; Franzini, D.; Harris, B.; Copeland, E. M., III. Neuroendocrine (Merkel cell) carcinoma of the skin. Its natural history, diagnosis, and treatment. Ann. Surg. 207:201-207; 1988. 11. Knox, S. J.; Kapp, D. S. Hyperthermia and radiation therapy in the treatment of recurrent Merkel cell tumors. Cancer 62:1479-1486; 1988. 12. Mackay, B.; Luna, M. A.: Butler, J. J. Adult neuroblastoma:
Merkel cell carcinoma 0 W. H. MORRISONet al.
13.
14.
15.
16.
17.
18.
electron microscopic observations in nine cases. Cancer 37: 1334-1351: 1976. Meland, N. B.; Jackson, 1. T. Merkel cell tumor: diagnosis, prognosis, and management. Plast. Reconstr. Surg. 77:632638: 1986. Merkel, F. Tastzellen und Tastkoerperchen bei den haushieren und beim menschen. Arch. Mikr. Anat. 11:636-652; 1875. O’Brien, P. C.; Denham, J. W.; Leong, S-Y. Merkel cell carcinoma: a review of behaviour patterns and management strategies. Aust. N. Z. J. Surg. 57:847-850: 1987. Pacella, J. A.; Ashby, M. A. Response to editorial “Neuroendocrine carcinoma of the skin: diagnostic and management considerations” (Letter). Int. J. Radiat. Oncol. Biol. Phys. 15:249: 1988. Pacella, J. A.; Ashby, M. A.: Ainslie, J.; Minty, C. The role of radiotherapy in the management of primary cutaneous neuroendocrine tumors (Merkel cell or trabecular carcinoma): experience at the Peter MacCallum Cancer Institute (Melbourne, Australia). Int. J. Radiat. Oncol. Biol. Phys. 14:1077-1084; 1988. Pilotti, S.; Rilke. F.; Bartoli, C.: Grisotti, A. Clinicopathologic
19.
20.
2 1.
22. 23. 24. 25. 26.
591
correlations of cutaneous neuroendocrine Merkel cell carcinoma. J. Clin. Oncol. 6: 1863-1873; 1988. Sibley, R. K.; Dehner, L. P.; Rosai, J. Primary neuroendocrine (Merkel cell?) carcinoma of the skin. I. A clinicopathologic and ultrastructural study of 43 cases. Am. J. Surg. Pathol. 9:95-108; 1985. Silva, E.; Mackay, B. Neuroendocrine (Merkel cell) carcinomas of the skin: an ultrastructural study of nine cases. Ultrastruct. Pathol. 2: l-9; 198 1. Silva, E. G.; Mackay, B.; Goepfert, H.; Burgess, M. A.; Fields, R. S. Endocrine carcinoma of the skin (Merkel cell carcinoma). Pathol. Annu. l9:1-30; 1984. Tang, C-K.: Toker, C. Trabecular carcinoma of the skin. An ultrastructural study. Cancer 42:23 I l-232 I ; 1978. Tapley, N. duV. Clinical applications ofthe electron beam. New York: John Wiley & sons; 1976. Taxy, J. B.; Ettinger, D. S.; Wharam, M. D. Primary small cell carcinoma of the skin. Cancer 46:2308-231 I: 1980. Toker. C. Trabecular carcinoma of the skin. Arch. Dermatol. 105:107-l IO; 1972. Wynne, C. J.: Kearsley, J. H. Merkel cell tumor: a chemosensitive skin cancer. Cancer 62:28-3 1; 1988.