- Email: [email protected]
Clinical staging of rectal cancer
Clinical Staging of Rectal Cancer
Jerome
S. Abrams,
MD, Burlington,
Vermont
Failure to improve the survival rate in patients with rectal cancers by radical surgery, with or without adjuvant radiotherapy or chemotherapy, has led some surgeons to recommend more conservative procedures such as local excision, electrocoagulation or endorectal radiation. Advocates of these procedures have stressed the importance of careful patient selection based largely on clinical staging, that is, 1tGons less t.han 5 cm in greatest, diameter with no clsntral ulceration. In addition, most select only patients with histologically differentiated lesions. A survey of 612 patients with 649 cancers of the colon and rectum undergoing elective resection in Vermont during the period 1971 to 1975 was recently completed [I]. One hundred sixty-seven lesions (26 percent) located 15 cm or less from the anorectal junction were classified as rectal cancers. The present sl,udy evaluates the validity of clinical staging by assessing the degree of correlation between size, morphology and degree of differentiation of each tumor with the histologic Dukes’ classification. Material
and Methods
The record of each patient with a rectal cancer was reviewed and the following information obtained: (1) maximum diameter of the tumor, (2) gross morphology, that is,
ulceration or nonulceration as recorded by the pathologist during gross examination of the specimen, (3) the microscopic degree of differentiation of the primary tumor, (4) the extent of histologic penetration of the tumor through the bowel wall or into perirectal tissues, (5) assessment of all lymph nodes in the specimen for the presence of metastatic disease, and (6) when available, histologic evidence 01’extension of cancer into contiguous organs or metastasis tc. the liver. From the Department of Surgery, College of Medicine, The University of Vermont, Burlington, Vermont. Presented at the 60th Annual Meeting of the New England Surgical Society, Whitefield. New Hampshire, September 28-30. 1979. Reprint requests should be addressed to Jerome S. Abrams, MD, Department of Surgery, College of Medicine, The University of Vermont, Burlington, Vermont 05405.
Volume 139, April 1980
All clinical and histologic staging followed the classification system proposed by Dukes [z] in 1932: An A cancer has penetrated the muscularis mucosa into the submucosa or into but not completely through the muscularis propria; when the tumor has penetrated through all layers of the muscularis propria to the serosa or into perirectal fat, it is classified B; any lesion associated with one or more positive lymph nodes, irrespective of the extent of penetration through the bowel wall, or that has spread to a contiguous organ or metastasized to liver or chest is classified Dukes’ C or D. The clinical impression of liver metastases at laparotomy was usually not substantiated by biopsy, and clinically contiguous spread of tumor was rarely documented by histologic examination. Carcinoma that did not invade through the muscularis mucosa into the submucosa was classified as “carcinoma in situ”; these cases were excluded from the study. Results
Fifty-three or 32 percent of the 167 rectal cancers were classified histologically as Dukes’ A. The correlation between size and histologic classification is shown in Table I. Although a larger proportion of the tumors with a diameter of less than 3 cm were in class A than those with a diameter of 3 to 5 cm (44 percent versus 33 percent), this difference is not statistically significant. The difference becomes significant, however, when cancers that were less than 5 cm are compared with larger lesions, only 16 percent of which were classified as Dukes’ A. The presence or absence of ulceration was noted in 99 of the 115 cancers measuring 5 cm or less. Sixty-three percent of nonulcerated cancers but only 28 percent of ulcerated lesions were classified as Dukes’ A; this difference is highly significant (p <0.005). Of the 30 nonulcerated cancers, 13 percent were class B and 23 percent class C or D. Of the 69 ulcerated cancers, 38 percent were class B and 39 percent class C or D. The pathologic report on 81 of the 115 cancers that measured 5 cm or less included information regarding
539
Abrams
TABLE I
Correlation Between Tumor Size (Maximum Diameter) and Histologic Dukes’ Classification in 167 Rectal Cancers
Size
(cm)
No. of Cancers
A
13 3-5 15 >5
48 67 115 37
44 33 37 16’
l
Dukes’ Class B 25 25 25 46
(%) CandD 31 42 38 38
p G.05.
70
70
g 60 t 2 50 "
60
a 40 *ul E 30 2 bQ 20
40
20
IO
10
50
30
n
the degree of differentiation of the tumor. Twentyfive of 54 well differentiated or partially differentiated lesions (46 percent) but only 5 of 27 undifferentiated cancers (19 percent) were classified as Dukes’ A (p <0.05). These data are summarized in Figure 1. Comments In 1977 Greaney and Irvin [3] reviewed 52 rectal cancers examined after abdominoperineal resection. Thirteen lesions measured less than 5 cm in diameter (25 percent), compared with 27 percent in the present study; 25 percent were nonulcerated, compared with 30 percent in the Vermont series. Fifty-six percent of the nonulcerated lesions proved to be histologic Dukes’ A, compared with 63 percent in the present study, whereas only 29 percent of ulcerating lesions were classified as A, compared with 28 percent in the Vermont series. The remarkable similarity between the data obtained from two series of patients with carcinoma of the rectum leaves little doubt as to the clinician’s ability to identify Dukes’ A lesions with 60 to 70 percent accuracy using biopsy and histologic examination to rule out poorly differentiated neoplasms. The major criticism of the validity of this observation concerns the histologic interpretation of the extent of penetration through the bowel wall. As Rubio et al [4] have noted, while routine microscopic sections obtained from 23 resected cancers showed that 7 were histologically Dukes’ A, when additional sections were reviewed, 6 of the 7 showed invasion through all layers of the muscularis propria and were reclassified as B lesions. They recommended that whenever there was invasion of the muscularis propria on routine section without penetration through all layers, additional sections should be reviewed. It must be assumed, therefore, that some of the cancers classified as Dukes’ A in the present series were really in class B and that some specimens may have contained
540
I
II
III
IV
v
n
-
‘p < .05 Figure 7. incidence of Dukes’ A cancers in relafton fo the size, morphology and hfstokgk differentiafton of the lesions. Column I represents all 167 recta/ cancers, column II, I?5 cancers 5 cm or less in diameter, column Ill, 54 lesions with a diameter of 5 cm or less that were partly or welt differentiated, column IV, 30 ksions measuring 5 cm or less that were not ulcerated and column V, 20 nonukerated, histologically differentiated cancers with a diameter of 5 cm or less.
positive lymph nodes that were not detected by the pathologist. As long as surgical treatment consists of radical removal of the rectum by low anterior or abdominoperineal resection, clinical staging is largely academic. It is the renewed interest in lesser procedures, based largely on clinical staging, that necessitates critical reevaluation of the surgeon’s ability to stage the disease without having a specimen to examine under the microscope. In 1978 Lock et al [5] at the St. Mark’s Hospital reported on a series of patients who underwent local excision of cancers of the colon and rectum during the 30 year period 1948 through 1977. Twenty-two patients had nonpedunculated cancers of the rectum that would have been managed by abdominoperineal resection; the general indications for local excision in these patients were advanced age, poor general health or obesity. Eleven of the 22 lesions (50 percent) were nonulcerated, and 14 (64 percent) had a diameter of 3 cm or less. The lesions were removed by a “full thickness disc excision with careful histologic examination of all margins of the excised cancer.” Five of the 22 patients (23 percent) underwent subsequent resection because of concern regarding the completeness of the primary excision or evidence of persistent tumor. At 5 years, 14 of 15 patients whose management was limited to local excision were alive without evidence of recurrence. Thus, 14 of the
The American Journal 01 Surgery
Clinical Staging of Rectal Cancer
TABLE
II
Expected 5 Year Survival Radical Resection
Rate for 167 Patients
No. of Patients Dukes’ A Dukes’ B Dukes’ C and D Total Operative mortality Alive at 5 years Gain
With Rectal
Cancers:
Conservative
Treatment
~__
_~._~.
__Survival Rate (%)’
45 60 62 167
90 60 13
‘;;
Versus
Conservative Treatment 41 t t 41 (25%) &‘(25%)
_._._____-
Radical Resection 41 36 6 85 (51%) -12 73 (44%) +32 (19%)
* Based on the author’s data (1971-1975). Assumes no survival with Dukes’ B. C and D.
22 patients (64 percent) were essentially “cured” of Dukes’ A cancers. All five of the patients who underwent. stlhsequent radical surgery were also “cllred,” and of t>wo patient,s who developed late recuerence and were treated with repeat local excision, one was alive wit,h cancer while the other was free of dit,ease. Only :< of 22 patients had died or had evidence of’ recurrent cancer. None of the lesions included in their series was of “high grade malignancy”; that is, all showed significant, evidence of different iation. Wadden and Kendalaft [h‘] reported their experien1.e with 77 patients with rectal cancer treated by electrocoagnlation. Fifty patients (65 percent) were under age 70 years, suggesting that age alone was not a rna,ior factor in case selection. According to the au1 hors, however, almost one third of their patients had been caonsidered unsuitahle candidates for abdominoperineal resection. Sixty-two percent of the lesions were less than 5 cm in diameter. Seventeen pat.ients had circumferential or villous lesions. Of the remaining 60 lesions, 35 or 57 percent were polypoid (nc~~nulcerated). On the basis of a 4 year follow-up, 21 ol’ the 34 patients with nonulcerated cancers (62 percent) versus 12 of 26 patients (46 percent) with ulcerating lesions were alive with no evidence of recurrence. The overall 4 year survival rate in these 60 patients was 55 percent. There were no postoperative de;lth:. hut the complications included two perforat Ions, hemorrhage and rect,ovaginal fistula in 2 of :i:< ‘Women. FSvcn more f’avorable results were reported by Crile i\nc ‘1’1~~nhull [ 71. In their series of 62 patients treated hy I~lec,t,rc,coag~llation, the average age was 67 years; 8 patirtl t 5 or 13 percent were over age 80 years. Ohvimsly. age played a role in selection since in the Vermont survey of’ 167 rectal cancers treated by resection only :i percent of patients were octogenarians 1 I 1. ‘I’he average size of the coagulated cancers was
Volume 139, April 1980
3.1 cm. By comparison, the average size of the lesion removed in 226 patients undergoing ahdominoperineal resection at their clinic was 4.8 cm, suggesting that size, too, played a role in selection. Twent,y-seven or 44 percent of the tumors t,reated by elertrocoagulation were nonulcerated. There were no deaths in this series of 62 patients, compared with nine postoperative deaths (4 percent) in the 226 patients undergoing abdominoperineal resection. At, 5 years, 68 percent of patients treated by electrocoagulation were alive, versus only 46 percent of’ patients undergoing ahdominoperineal resection. Twelve of the patients treated by coagulation also received radiotherapy. The higher survival rates reported hy Turnhull and Crile may represent a somewhat higher degree of selectivity (smaller tumors) than that used by Madden and Kendalaft [h’]. Excellent results were reported by Papillon [8] in 1975 in 133 patients treated by endocavitary radiation. The criteria for case selection included: (1) the tumor was well differentiated on histologic examination of a hiopsied specimen; (2) it lay below 12 cm from the anorectal juncti’on, (3) t,he tumor was “limited to the bowel wall” on digital examination; (4) the size did not exceed 5 by 3 cm, and (5) most tumors were described as polypoid. One hundred one of the IX] patients (76 percent) were alive after 5 years without evidence of recurrence. Only three required a subsequent ahdominoperineal resection for recurrent tumor. Unfortunately, some authors have heen tempted to compare the zero mortality and high apparent cure rates after these lesser procedures with results from abdominoperineal resection, ignoring the difference in patient selection. Table II shows the potential survival in the 167 patients reported on herein after radical resection versus treatment hy nonresectional methods. The following assumptions were made: (1) conservative procedures could cure Dukes A cancers
541
Abrams
TABLE III
Cancers All 55 cm 15 cm nonulcerating
Expected 5 Year Survival Rate for Patients With Rectal Cancer: Conservative Treatment Versus Radical Resection Depending on Case Selection
No. 167 115 30
Conservative Treatment No. % 41 38 17
25’ 33’ 57’
Radical Resection No. % 73 53 ia
44+ 46+ 60+
Gain 32 15 1
mended that efforts be made to institute a controlled prospective study. While radical resection remains the standard for the treatment of carcinoma of the rectum, there is sufficient evidence to recommend a more conservative procedure in poor risk patients (those with advanced age, senility, obesity or debilitating systemic disease) whose lesions meet the criteria of size, morphology and differentiation. Summary
* Assumes no survival with Dukes’ 8, C and D. + Assumes 7 percent operative mortality.
to the same degree as radical resection but were ineffective for B, C or D lesions; (2) the 5 year survival rate was based on data on 58 patients undergoing radical resection for cancer of the rectum reported by Abrams and Reines [9]; (3) radical resection was followed by an operative mortality rate of 7 percent [I]. Without patient selection, the expected 5 year survival rate after a lesser procedure would be 25 percent. Since 12 postoperative deaths would occur in the 167 patients treated by radical resection, the number of 5 year survivors would be reduced from 85 to 73, or 44 percent. This would still represent a gain of 32 patients over the series treated conservatively. As shown in Table III, limiting the selection of patients for lesser procedures to those with lesions less than 5 cm in diameter would increase the 5 year survival rate to 33 percent and reduce the gain from radical resection to 15 patients. If selection is further limited to nonulcerated cancers, the expected 5 year survival rate increases to 57 percent and the apparent gain from radical surgery is only one patient. These data are similar to those reported by Madden and Kendalaft [6], C!?ile and Turnbull [7] and Papillon
ISI. A randomized trial comparing the results of local excision, electrocoagulation and endocavitary radiation versus radical resection in selected patients meeting the criteria of size, morphology and degree of differentiation is long overdue. Two major problems exist: (1) only 15 to 20 percent of all rectal cancers will meet the criteria, and thus multiple institutions will have to participate in such a study, and (2) all of the lesser techniques require care and precision. As Wanebo and Quan [lo] noted in their follow-up of 18 patients who had developed recurrence after electrocoagulation, 7 procedures had been performed in the surgeon’s office. Both problems can be overcome, however, and it is strongly l’ecom-
542
The gross and microscopic pathologic features of 167 rectal cancers were reviewed and the size, morphology and degree of differentiation correlated with the histologic Dukes’ classification. A diagnosis of Dukes’ A adenocarcinoma was made in 32 percent of all rectal cancers and in 70 percent of nonulcerated, partly or well differentiated cancers that did not exceed 5 cm in greatest diameter. The incidence of Dukes’ A lesions in patients meeting these criteria is similar to the survival rates reported after local excision, electrocoagulation and endorectal radiation. A prospective study comparing the results of local treatment with radical resection in selected patients is recommended. References 1. Abrams JS. Elective resection for colorectal cancer in Vermont: 1971-1975. Am J Surg i980;139:78-83. 2. Dukes CE. The classification of cancer of the rectum. J Pathol Bacterial 1932;35:323-32. 3. Greaney MG, Irvin TT. Criteria for the selection of rectal cancers for local treatment. Dis Colon Rectum 1977;20:463-6. 4. Rubio CA, Emis S, Nylander G. A critical reappraisal of Dukes’ classification. Surg Gynecol Obstet 1977;145:662-5. 5. Lock MR, Carius DW, Ritchie JK, Lockhart-Mummary HE. The treatment of early colorectal cancer by local excision. Br J Surg 197&65:346-g. 6. Madden JL, Kandalaft S. Electrocoagulation in the treatment of cancer of the rectum. Ann Surg 1971;174:530-40. 7. Crile GC Jr, Turnbull RB. The role of the electrocoagulation in the treatment of carcinoma of the rectum. Surg Gynecol Obstet 1972;135:391-6. 6. Papillon J. Resectable rectal cancers. Treatment by curative endocavitary irradiation. JAMA 1975;231:1365-7. 9. Abrams JS, Reines HD. Increasing incidence of right-sided lesions in colorectal cancer. Am J Surg 1979;137:522-6. 10. Wanebo JJ, Quan SHQ. Failures of electrocoagulation of primary carcinoma of the rectum. Surg Gynecol Obstet 1974;136:174-6.
Discussion Charles S. Whelan (Worcester, MA): Dr. Abrams does well to turn our attention to the more conservative ways of managing rectal cancer. This is in line with the present trend away from the more radical operations for cancer. As Dr. Abrams mentions, lesser procedures include local
The American Journal of Surgery
Clinical
rxiision. rle~trocoaglllation and endorectal irradiation. I feel that irradiation has a limited application, being suited to smaller exophytic lesions. There does not seem to be a neat way of doing local excisional surgery for rectal ca ricer. Having had early experience in electrosurgery under the late T)r. (‘lif’f’ord Franseen and personal guidance by Dr. Madden, I use electrocoagulation as the procedure of choice for rectal cancer. Bulky tumors invading the whole bowel w;ll can be effectively treated. The limitations are the height. of the lesion (not more than 10 cm) and a tumor that is I,irc.~imt.erenti;ll. ‘I’hiit elt~c~trocoagulation is an effective method is amply pr<,ven hy kladden‘s excellent results in about 200 cases. M.s experience comprises 11 cases of rectal cancer so treated. The results have been similar. The patients un-
Volume
139, April 1990
Staging
of Rectal
Cancer
derwent a procedure that results in little or no mortality and none of the urinary or genital tract complications of abdominoperineal resection, and they go on living without the burden of a colostomy. The Dukes’ classification does not apply when one uses this approach. One simply persists in coagulat,ion until all tumor is removed. The proper depth comes with experience. It is regrettable that this technique is not being taught in our medical centers; every area should have a few surgeons who have mastered it. Jerome S. Abrams (closing): I would like to t,hank Dr. Whelan for his remarks and say again that I am not proposing any of the local procedures that have been recommended, but rather suggesting that with careful selection, one can to a great degree select favorable rectal cancers.
543
Jerome
S. Abrams,
MD, Burlington,
Vermont
Failure to improve the survival rate in patients with rectal cancers by radical surgery, with or without adjuvant radiotherapy or chemotherapy, has led some surgeons to recommend more conservative procedures such as local excision, electrocoagulation or endorectal radiation. Advocates of these procedures have stressed the importance of careful patient selection based largely on clinical staging, that is, 1tGons less t.han 5 cm in greatest, diameter with no clsntral ulceration. In addition, most select only patients with histologically differentiated lesions. A survey of 612 patients with 649 cancers of the colon and rectum undergoing elective resection in Vermont during the period 1971 to 1975 was recently completed [I]. One hundred sixty-seven lesions (26 percent) located 15 cm or less from the anorectal junction were classified as rectal cancers. The present sl,udy evaluates the validity of clinical staging by assessing the degree of correlation between size, morphology and degree of differentiation of each tumor with the histologic Dukes’ classification. Material
and Methods
The record of each patient with a rectal cancer was reviewed and the following information obtained: (1) maximum diameter of the tumor, (2) gross morphology, that is,
ulceration or nonulceration as recorded by the pathologist during gross examination of the specimen, (3) the microscopic degree of differentiation of the primary tumor, (4) the extent of histologic penetration of the tumor through the bowel wall or into perirectal tissues, (5) assessment of all lymph nodes in the specimen for the presence of metastatic disease, and (6) when available, histologic evidence 01’extension of cancer into contiguous organs or metastasis tc. the liver. From the Department of Surgery, College of Medicine, The University of Vermont, Burlington, Vermont. Presented at the 60th Annual Meeting of the New England Surgical Society, Whitefield. New Hampshire, September 28-30. 1979. Reprint requests should be addressed to Jerome S. Abrams, MD, Department of Surgery, College of Medicine, The University of Vermont, Burlington, Vermont 05405.
Volume 139, April 1980
All clinical and histologic staging followed the classification system proposed by Dukes [z] in 1932: An A cancer has penetrated the muscularis mucosa into the submucosa or into but not completely through the muscularis propria; when the tumor has penetrated through all layers of the muscularis propria to the serosa or into perirectal fat, it is classified B; any lesion associated with one or more positive lymph nodes, irrespective of the extent of penetration through the bowel wall, or that has spread to a contiguous organ or metastasized to liver or chest is classified Dukes’ C or D. The clinical impression of liver metastases at laparotomy was usually not substantiated by biopsy, and clinically contiguous spread of tumor was rarely documented by histologic examination. Carcinoma that did not invade through the muscularis mucosa into the submucosa was classified as “carcinoma in situ”; these cases were excluded from the study. Results
Fifty-three or 32 percent of the 167 rectal cancers were classified histologically as Dukes’ A. The correlation between size and histologic classification is shown in Table I. Although a larger proportion of the tumors with a diameter of less than 3 cm were in class A than those with a diameter of 3 to 5 cm (44 percent versus 33 percent), this difference is not statistically significant. The difference becomes significant, however, when cancers that were less than 5 cm are compared with larger lesions, only 16 percent of which were classified as Dukes’ A. The presence or absence of ulceration was noted in 99 of the 115 cancers measuring 5 cm or less. Sixty-three percent of nonulcerated cancers but only 28 percent of ulcerated lesions were classified as Dukes’ A; this difference is highly significant (p <0.005). Of the 30 nonulcerated cancers, 13 percent were class B and 23 percent class C or D. Of the 69 ulcerated cancers, 38 percent were class B and 39 percent class C or D. The pathologic report on 81 of the 115 cancers that measured 5 cm or less included information regarding
539
Abrams
TABLE I
Correlation Between Tumor Size (Maximum Diameter) and Histologic Dukes’ Classification in 167 Rectal Cancers
Size
(cm)
No. of Cancers
A
13 3-5 15 >5
48 67 115 37
44 33 37 16’
l
Dukes’ Class B 25 25 25 46
(%) CandD 31 42 38 38
p G.05.
70
70
g 60 t 2 50 "
60
a 40 *ul E 30 2 bQ 20
40
20
IO
10
50
30
n
the degree of differentiation of the tumor. Twentyfive of 54 well differentiated or partially differentiated lesions (46 percent) but only 5 of 27 undifferentiated cancers (19 percent) were classified as Dukes’ A (p <0.05). These data are summarized in Figure 1. Comments In 1977 Greaney and Irvin [3] reviewed 52 rectal cancers examined after abdominoperineal resection. Thirteen lesions measured less than 5 cm in diameter (25 percent), compared with 27 percent in the present study; 25 percent were nonulcerated, compared with 30 percent in the Vermont series. Fifty-six percent of the nonulcerated lesions proved to be histologic Dukes’ A, compared with 63 percent in the present study, whereas only 29 percent of ulcerating lesions were classified as A, compared with 28 percent in the Vermont series. The remarkable similarity between the data obtained from two series of patients with carcinoma of the rectum leaves little doubt as to the clinician’s ability to identify Dukes’ A lesions with 60 to 70 percent accuracy using biopsy and histologic examination to rule out poorly differentiated neoplasms. The major criticism of the validity of this observation concerns the histologic interpretation of the extent of penetration through the bowel wall. As Rubio et al [4] have noted, while routine microscopic sections obtained from 23 resected cancers showed that 7 were histologically Dukes’ A, when additional sections were reviewed, 6 of the 7 showed invasion through all layers of the muscularis propria and were reclassified as B lesions. They recommended that whenever there was invasion of the muscularis propria on routine section without penetration through all layers, additional sections should be reviewed. It must be assumed, therefore, that some of the cancers classified as Dukes’ A in the present series were really in class B and that some specimens may have contained
540
I
II
III
IV
v
n
-
‘p < .05 Figure 7. incidence of Dukes’ A cancers in relafton fo the size, morphology and hfstokgk differentiafton of the lesions. Column I represents all 167 recta/ cancers, column II, I?5 cancers 5 cm or less in diameter, column Ill, 54 lesions with a diameter of 5 cm or less that were partly or welt differentiated, column IV, 30 ksions measuring 5 cm or less that were not ulcerated and column V, 20 nonukerated, histologically differentiated cancers with a diameter of 5 cm or less.
positive lymph nodes that were not detected by the pathologist. As long as surgical treatment consists of radical removal of the rectum by low anterior or abdominoperineal resection, clinical staging is largely academic. It is the renewed interest in lesser procedures, based largely on clinical staging, that necessitates critical reevaluation of the surgeon’s ability to stage the disease without having a specimen to examine under the microscope. In 1978 Lock et al [5] at the St. Mark’s Hospital reported on a series of patients who underwent local excision of cancers of the colon and rectum during the 30 year period 1948 through 1977. Twenty-two patients had nonpedunculated cancers of the rectum that would have been managed by abdominoperineal resection; the general indications for local excision in these patients were advanced age, poor general health or obesity. Eleven of the 22 lesions (50 percent) were nonulcerated, and 14 (64 percent) had a diameter of 3 cm or less. The lesions were removed by a “full thickness disc excision with careful histologic examination of all margins of the excised cancer.” Five of the 22 patients (23 percent) underwent subsequent resection because of concern regarding the completeness of the primary excision or evidence of persistent tumor. At 5 years, 14 of 15 patients whose management was limited to local excision were alive without evidence of recurrence. Thus, 14 of the
The American Journal 01 Surgery
Clinical Staging of Rectal Cancer
TABLE
II
Expected 5 Year Survival Radical Resection
Rate for 167 Patients
No. of Patients Dukes’ A Dukes’ B Dukes’ C and D Total Operative mortality Alive at 5 years Gain
With Rectal
Cancers:
Conservative
Treatment
~__
_~._~.
__Survival Rate (%)’
45 60 62 167
90 60 13
‘;;
Versus
Conservative Treatment 41 t t 41 (25%) &‘(25%)
_._._____-
Radical Resection 41 36 6 85 (51%) -12 73 (44%) +32 (19%)
* Based on the author’s data (1971-1975). Assumes no survival with Dukes’ B. C and D.
22 patients (64 percent) were essentially “cured” of Dukes’ A cancers. All five of the patients who underwent. stlhsequent radical surgery were also “cllred,” and of t>wo patient,s who developed late recuerence and were treated with repeat local excision, one was alive wit,h cancer while the other was free of dit,ease. Only :< of 22 patients had died or had evidence of’ recurrent cancer. None of the lesions included in their series was of “high grade malignancy”; that is, all showed significant, evidence of different iation. Wadden and Kendalaft [h‘] reported their experien1.e with 77 patients with rectal cancer treated by electrocoagnlation. Fifty patients (65 percent) were under age 70 years, suggesting that age alone was not a rna,ior factor in case selection. According to the au1 hors, however, almost one third of their patients had been caonsidered unsuitahle candidates for abdominoperineal resection. Sixty-two percent of the lesions were less than 5 cm in diameter. Seventeen pat.ients had circumferential or villous lesions. Of the remaining 60 lesions, 35 or 57 percent were polypoid (nc~~nulcerated). On the basis of a 4 year follow-up, 21 ol’ the 34 patients with nonulcerated cancers (62 percent) versus 12 of 26 patients (46 percent) with ulcerating lesions were alive with no evidence of recurrence. The overall 4 year survival rate in these 60 patients was 55 percent. There were no postoperative de;lth:. hut the complications included two perforat Ions, hemorrhage and rect,ovaginal fistula in 2 of :i:< ‘Women. FSvcn more f’avorable results were reported by Crile i\nc ‘1’1~~nhull [ 71. In their series of 62 patients treated hy I~lec,t,rc,coag~llation, the average age was 67 years; 8 patirtl t 5 or 13 percent were over age 80 years. Ohvimsly. age played a role in selection since in the Vermont survey of’ 167 rectal cancers treated by resection only :i percent of patients were octogenarians 1 I 1. ‘I’he average size of the coagulated cancers was
Volume 139, April 1980
3.1 cm. By comparison, the average size of the lesion removed in 226 patients undergoing ahdominoperineal resection at their clinic was 4.8 cm, suggesting that size, too, played a role in selection. Twent,y-seven or 44 percent of the tumors t,reated by elertrocoagulation were nonulcerated. There were no deaths in this series of 62 patients, compared with nine postoperative deaths (4 percent) in the 226 patients undergoing abdominoperineal resection. At, 5 years, 68 percent of patients treated by electrocoagulation were alive, versus only 46 percent of’ patients undergoing ahdominoperineal resection. Twelve of the patients treated by coagulation also received radiotherapy. The higher survival rates reported hy Turnhull and Crile may represent a somewhat higher degree of selectivity (smaller tumors) than that used by Madden and Kendalaft [h’]. Excellent results were reported by Papillon [8] in 1975 in 133 patients treated by endocavitary radiation. The criteria for case selection included: (1) the tumor was well differentiated on histologic examination of a hiopsied specimen; (2) it lay below 12 cm from the anorectal juncti’on, (3) t,he tumor was “limited to the bowel wall” on digital examination; (4) the size did not exceed 5 by 3 cm, and (5) most tumors were described as polypoid. One hundred one of the IX] patients (76 percent) were alive after 5 years without evidence of recurrence. Only three required a subsequent ahdominoperineal resection for recurrent tumor. Unfortunately, some authors have heen tempted to compare the zero mortality and high apparent cure rates after these lesser procedures with results from abdominoperineal resection, ignoring the difference in patient selection. Table II shows the potential survival in the 167 patients reported on herein after radical resection versus treatment hy nonresectional methods. The following assumptions were made: (1) conservative procedures could cure Dukes A cancers
541
Abrams
TABLE III
Cancers All 55 cm 15 cm nonulcerating
Expected 5 Year Survival Rate for Patients With Rectal Cancer: Conservative Treatment Versus Radical Resection Depending on Case Selection
No. 167 115 30
Conservative Treatment No. % 41 38 17
25’ 33’ 57’
Radical Resection No. % 73 53 ia
44+ 46+ 60+
Gain 32 15 1
mended that efforts be made to institute a controlled prospective study. While radical resection remains the standard for the treatment of carcinoma of the rectum, there is sufficient evidence to recommend a more conservative procedure in poor risk patients (those with advanced age, senility, obesity or debilitating systemic disease) whose lesions meet the criteria of size, morphology and differentiation. Summary
* Assumes no survival with Dukes’ 8, C and D. + Assumes 7 percent operative mortality.
to the same degree as radical resection but were ineffective for B, C or D lesions; (2) the 5 year survival rate was based on data on 58 patients undergoing radical resection for cancer of the rectum reported by Abrams and Reines [9]; (3) radical resection was followed by an operative mortality rate of 7 percent [I]. Without patient selection, the expected 5 year survival rate after a lesser procedure would be 25 percent. Since 12 postoperative deaths would occur in the 167 patients treated by radical resection, the number of 5 year survivors would be reduced from 85 to 73, or 44 percent. This would still represent a gain of 32 patients over the series treated conservatively. As shown in Table III, limiting the selection of patients for lesser procedures to those with lesions less than 5 cm in diameter would increase the 5 year survival rate to 33 percent and reduce the gain from radical resection to 15 patients. If selection is further limited to nonulcerated cancers, the expected 5 year survival rate increases to 57 percent and the apparent gain from radical surgery is only one patient. These data are similar to those reported by Madden and Kendalaft [6], C!?ile and Turnbull [7] and Papillon
ISI. A randomized trial comparing the results of local excision, electrocoagulation and endocavitary radiation versus radical resection in selected patients meeting the criteria of size, morphology and degree of differentiation is long overdue. Two major problems exist: (1) only 15 to 20 percent of all rectal cancers will meet the criteria, and thus multiple institutions will have to participate in such a study, and (2) all of the lesser techniques require care and precision. As Wanebo and Quan [lo] noted in their follow-up of 18 patients who had developed recurrence after electrocoagulation, 7 procedures had been performed in the surgeon’s office. Both problems can be overcome, however, and it is strongly l’ecom-
542
The gross and microscopic pathologic features of 167 rectal cancers were reviewed and the size, morphology and degree of differentiation correlated with the histologic Dukes’ classification. A diagnosis of Dukes’ A adenocarcinoma was made in 32 percent of all rectal cancers and in 70 percent of nonulcerated, partly or well differentiated cancers that did not exceed 5 cm in greatest diameter. The incidence of Dukes’ A lesions in patients meeting these criteria is similar to the survival rates reported after local excision, electrocoagulation and endorectal radiation. A prospective study comparing the results of local treatment with radical resection in selected patients is recommended. References 1. Abrams JS. Elective resection for colorectal cancer in Vermont: 1971-1975. Am J Surg i980;139:78-83. 2. Dukes CE. The classification of cancer of the rectum. J Pathol Bacterial 1932;35:323-32. 3. Greaney MG, Irvin TT. Criteria for the selection of rectal cancers for local treatment. Dis Colon Rectum 1977;20:463-6. 4. Rubio CA, Emis S, Nylander G. A critical reappraisal of Dukes’ classification. Surg Gynecol Obstet 1977;145:662-5. 5. Lock MR, Carius DW, Ritchie JK, Lockhart-Mummary HE. The treatment of early colorectal cancer by local excision. Br J Surg 197&65:346-g. 6. Madden JL, Kandalaft S. Electrocoagulation in the treatment of cancer of the rectum. Ann Surg 1971;174:530-40. 7. Crile GC Jr, Turnbull RB. The role of the electrocoagulation in the treatment of carcinoma of the rectum. Surg Gynecol Obstet 1972;135:391-6. 6. Papillon J. Resectable rectal cancers. Treatment by curative endocavitary irradiation. JAMA 1975;231:1365-7. 9. Abrams JS, Reines HD. Increasing incidence of right-sided lesions in colorectal cancer. Am J Surg 1979;137:522-6. 10. Wanebo JJ, Quan SHQ. Failures of electrocoagulation of primary carcinoma of the rectum. Surg Gynecol Obstet 1974;136:174-6.
Discussion Charles S. Whelan (Worcester, MA): Dr. Abrams does well to turn our attention to the more conservative ways of managing rectal cancer. This is in line with the present trend away from the more radical operations for cancer. As Dr. Abrams mentions, lesser procedures include local
The American Journal of Surgery
Clinical
rxiision. rle~trocoaglllation and endorectal irradiation. I feel that irradiation has a limited application, being suited to smaller exophytic lesions. There does not seem to be a neat way of doing local excisional surgery for rectal ca ricer. Having had early experience in electrosurgery under the late T)r. (‘lif’f’ord Franseen and personal guidance by Dr. Madden, I use electrocoagulation as the procedure of choice for rectal cancer. Bulky tumors invading the whole bowel w;ll can be effectively treated. The limitations are the height. of the lesion (not more than 10 cm) and a tumor that is I,irc.~imt.erenti;ll. ‘I’hiit elt~c~trocoagulation is an effective method is amply pr<,ven hy kladden‘s excellent results in about 200 cases. M.s experience comprises 11 cases of rectal cancer so treated. The results have been similar. The patients un-
Volume
139, April 1990
Staging
of Rectal
Cancer
derwent a procedure that results in little or no mortality and none of the urinary or genital tract complications of abdominoperineal resection, and they go on living without the burden of a colostomy. The Dukes’ classification does not apply when one uses this approach. One simply persists in coagulat,ion until all tumor is removed. The proper depth comes with experience. It is regrettable that this technique is not being taught in our medical centers; every area should have a few surgeons who have mastered it. Jerome S. Abrams (closing): I would like to t,hank Dr. Whelan for his remarks and say again that I am not proposing any of the local procedures that have been recommended, but rather suggesting that with careful selection, one can to a great degree select favorable rectal cancers.
543