222
1. J. Radiation
Oncology
0 Biology 0 Physics
could even be in a situation where we have less late effects at HDR than at LDR - for the same early effects. This phenomenon which, primafacie, appears contrary to basic radiobiological theory, depends on several factors, the main ones being the dose to the bladder/rectum relative to the treatment dose, and the relative curviness of the dose-response curves for early and late-responding tissues. This latter is usually described by “o/B” ratios; for the best available values of these parameters, we have shown that if the brachytherapy delivers less than about 75% of the treatment dose to the bladder/rectum, then if the HDR dose is reduced to give equal early effects to LDR, the HDR late effects should not be worse than the LDR late effects. Simply on these considerations, the smaller the number of fractions the better. In practice, however, other considerations, such as the need for tumor reoxygenation, would suggest use of no less than about five fractions. There is a third related factor to consider, which is that the short treatment time that is characteristic of HDR allows packing and retraction of the sensitive organs, estimated by Orton (4) to result in a 20% further decrease in rectal/bladder dose. This advantage results from an extra physical factor, on top of the biological factors discussed here. In light of these considerations, we would suggest that Orton’s suggestion (5) regarding the use of large numbers of fractions for HDR intracavitary brachytherapy of the cervix should be replaced by the following:
I. When the dose to the dose-limiting critical normal tissues (bladder/ rectum) is less than about 75% of the prescribed dose, for equal early response, HDR results in comparable (or less) late effects than LDR. 2. When HDR is indicated, about five fractions is probably optimal. There is no need for larger fraction numbers. 3. When HDR is indicated, HDR protocols comparable to those at LDR should be designed based on matching early rather than late effects. 4. For those patients where the dose to the bladder/rectum is comparable to the prescribed dose, HDR is contraindicated.
It should be noted, of course, that for other sites of interest for HDR (e.g., endometrial carcinoma), unless the dose-limiting organs receive a significantly smaller dose than the treatment dose, Orton’s comment is probably appropriate, and HDR might be expected to result in some loss in therapeutic efficacy. DAVIDJ. BRENNER,PH.D. ERIC J. HALL, D.Sc. Center for Radiological Research Columbia University 630 West 168th St. New York. NY 10032
1. Brenner, D. J.; Hall, E. J. Fractionated high dose-rate versus low dose-rate brachytherapy of the cervix. I. General considerations based on radiobiology. Brit. J. Radiol. 64:133-141; 1991. 2. Brenner, D. J.; Hall, E. J. Fractionated high dose-rate versus low dose-rate regimens for intracavitary brachytherapy of the cervix II. Equivalent regimes for combined brachytherapy and external irradiation. Int. J. Radiat. Oncol. Biol. Phys. (in press). 3. Dale, R. G. The use of small fraction numbers naecological afterloading: Some radiobiological J. Radiol. 63:290-294; 1990.
in high dose-rate gyconsiderations. Brit.
4. Orton, C. G. Remote Afterloading for cervix cancer: The physicist’s point of view. In: Martinez, A. A., Orton, C. G., Mould, R. F., eds. Brachytherapy HDR and LDR. The Netherlands: Nucletron; 1990. 5. Orton, C. G. HDR in gynecological applications: Dose and volume considerations. Int. J. Radiat. Oncol. Biol. Phys. 20:1379-1380; 1991. 6. Van Lancker, M.; Storme, G. Prediction of severe late complications in fractionated high dose rate brachytherapy in gynecological applications. Int. J. Radiat. Oncol. Biol. Phys. 20:1125-1129; 1990. 7. Withers, H. R.; Thames, H. D.; Peters, L. J. Differences in the fractionation response of acutely and late-responding tissues. In: Karcher,
Volume 22, Number
1. 1992
K. H., Kogelnik, H. D., Reinartz, G., eds. Progress in radio-oncology, vol. II. New York, NY: Raven Press; 1982:287-296.
FRACTIONATION IS IMPORTANT FOR HDR CERVIX CANCER BRACHYTHERAPY To the Editor: Drs. Brenner and Hall (2) make a very interesting observation concerning HDR brachytherapy when used for treatments when the dose to the limiting late-reacting normal tissues is below a certain “threshold” fraction of that to the tumor. Under these conditions, doses to normal tissues are so low, that increasing the number of HDR fractions no longer increases the therapeutic ratio. This is because cell survival curves for tumors and late-reacting normal tissues have different shapes (see their Fig. 1) and this leads to more sparing of tumor cells than normal cells by increasing fractionation. Up to this point, I completely agree with Drs. Brenner and Hall. But I completely disagree with them on the normal-tissue threshold dose below which this phenomenon occurs, and the subsequent relevance to cervix cancer radiotherapy. In their Letter to the Editor (2), they state that if the dose to rectal and bladder tissues is less than about 75% of the Point A dose, fractionation of the HDR treatments decreases the therapeutic ratio, that is, the threshold fraction is 0.75. They also state that Dale (4) agrees with them. However, they fail to report that, although Dale does agree that if the ratio of the rectal/bladder to Point A doses, which he calls the geometrical sparing factor m, is below a certain critical value, then fractionation decreases the therapeutic ratio; he shows graphically (Dale’s Fig. I) that this threshold value off is about 0.3, not 0.75. This is a vitally important difference, since average “hot-spot” rectal/bladder doses in HDR brachytherapy are typically 0.64.7 of the Point A doses (6). If Dale is correct, then increased fractionation is good, whereas if Brenner and Hall are right, it is not. Since both the Dale (4) and Brenner and Hall (2) studies both apply the linear-quadratic (L-Q) model to arrive at their conclusions, this difference is not due to the models they utilize but, as shown in the Appendix, is caused by the parameters (a/p values) used in these models. As shown in the Appendix, the critical threshold value off is given by:
where (u/B), and (o/B), are the a/B values for late-reacting normal tissues and tumor, respectively. Dale (4) states that he has used (a/B), = 3 Gy and (o/B), = IO Gy and this leads to a threshold value offof 0.3. Brenner and Hall do not state in their letter what a/B values they have assumed, but reference to an earlier paper (1) shows that they have probably used (n/B), = 4 Gy and (o/B), = 5.3 Gy, which leads to their threshold value offof 0.75. To return now to the original question: Does increased fractionation of HDR brachytherapy for cervix cancer improve the therapeutic ratio? According to the predictions of the L-Q model the answer is clearly yes, if one accepts the typical a/B values published for late-reacting tissues and tumors of 3 Gy and IO Gy, respectively, since their ratio is 0.3, which is well below the observed f values of 0.60.7 (6). However, since we really do not know these a/B values with any confidence, and since the L-Q model itself is probably only an approximation, the ultimate answer to this question has to come from clinical experience. Fortunately, we now have the benefit of the results of a recent survey of 56 institutions, treating over 17,000 Ca cervix patients with HDR brachytherapy (6), which shows that increased fractionation plays a very significant role in decreasing the risk of late complications. For example, for Pt. A doses 5 7 Gy/fraction the risk of severe late complications is observed to be 2-3 times lower than for doses > 7 Gyi fraction (6). In conclusion, therefore, I adhere to my previous contention that, for cervix cancer HDR treatments it is important to “keep the dose per,fraction down by giving manyfractions” (5). COLING. GRTON, PH.D. Gershenson Radiation Oncology Center Harper Hospital and Wayne State University 3990 John R. Street Detroit, MI 48201
223
Correspondence
APPENDIX Question:
for a fixed bioeffect dose to the tumor, does the bioeffect dose to normal late-reacting tissues increase or decrease with increased fractionation?
Answer:
the L-Q equations for bioeffect doses (ERDs) for tumor (subscript I) and late-reacting tissues (subscript I) are (3):
ERD, = Nd,(l + &)
(1)
ERD, = Nfd, (1 + &)
(2)
where d, is the dose/fraction to the tumor delivered in N fractions, and f is the geometrical sparing factor (4), that is, the ratio of the normal tissue to tumor doses. Substituting the value of N from eq( 1) and eq(2) yields the following equation for ERD, in terms of d,:
ERD, =
f (ERD,) (1 + A)-’
(1 + &)
Then the rate of change of ERD, with d, is obtained by differentiating ERD,, with respect to d,, realizing that ERD, is to be kept constant. This differentiation yields the equation:
AERD,)
WERD,) _
d(d,)-
Clearly,
pm, -
1. Where reconstruction of the surgically excised area is such as to place the histologically incomplete margin beneath transported skin, either as a skin graft or rotation flap. Recurrence is, thus, only detected at a time when it has grown through the overlying tissue. 2. Where the pathology report identifies the presence of perineural extension as a component of the incomplete margin. Although this is uncommon for BCC’s, it still occurs and, if untreated, can lead to quite significant problems later on. 3. Where the BCC abuts up to particular anatomical areas, such as the nasolabial groove, medial canthus, or post-auricular groove. In these areas, subsequent evidence of disease can, in fact, be quite bulky and distant to the initial site. This relates to the ability of BCC’s to extend along, but not destroy, tissue planes. The above comments are not meant to imply that all patients necessarily need to be treated, but the policy of observation with a histologically incomplete margin must be put in the appropriate clinical context. DR. ROBERT SMEE Department of Radiation Oncology Prince of Wales Hospital Randwick, NSW 2031
WP),l
(43, b/P), (1 +
ERD, increases
with increase
in d, (or decrease
A&B), - (a/B), is positive,
Hence, the threshold given by:
Dr. F. F. Liu et al. (1) They have endeavoured to establish a management approach to a frequently encountered clinical situation. In doing so, they are further refining the natural history of basal cell carcinomas (BCC’s) of the skin. However, I would view with concern the broad generalization of their article that observation can be adopted for all BCC’s incompletely excised. This policy can certainly apply in situations in which follow-up is available and adequate anatomical assessment is assured. There are, however, situations in which observation alone is clinically inappropriate. These circumstances are:
(cllp)I I
that is, f > cdpj
value off above which fractionation
f rhwshoid
1. Liu, F.-F.; Maki, E.; Warde, P.; Payne, D.; Fitzpatrick, P. A management approach to incompletely excised basal cell carcinoma of the skin. Int. J. Radiat. Oncol. Biol. Phys. 20:423-428; 1991.
in N) when
is desirable
is
_ (@)I (cl/P),
1 Brenner, D. J.; Hall, E. J. Fractionated high dose-rate versus low dose-rate brachytherapy of the cervix. I. General considerations based on radiobiology. Brit. J. Radiol. 64: 133-141; 1991. 2. Brenner, D. J.; Hall, E. J. HDR brachytherapy for carcinoma of the cervix: Fractionation considerations. Int. J. Radiat. Oncol. Biol. Physics. (in press). 3. Dale, R. G. The application of the linear-quadratic dose-effect equation to fractionated and protracted radiotherapy. Brit. J. Radiol. 58: 515-528; 1985. 4. Dale, R. G. The use of small fraction numbers naecological afterloading: Some radiobiological J. Radiol. 63:290-294; 1990.
in high dose-rate gyconsiderations. Brit.
5. Orton, C. G. HDR in gynecological applications: Dose and volume considerations. Int. J. Radiat. Oncol. Biol. Phys. 20:1379-1380; 1991. 6. Orton, C. G.; Seyedsadr, M.; Somnay, A. Comparison of high and low dose rate remote afterloading for cervix cancer and the importance of fractionation. Int. J. Radiat. Oncol. Biol. Physics 21:1425-1434; 1991.
CRITIQUE OF ‘A MANAGEMENT APPROACH TO INCOMPLETELY EXCISED BASAL CELL CARCINOMA OF THE SKIN’ To the Editor: I read with interest the recent article “A Management Approach to Incompletely Excised Basal Cell Carcinoma of the Skin” by
RESPONSE
TO DR. SMEE
To rhe Ediror: We thank Dr. R. Smee for his letter commenting on our paper “A Management Approach to Incompletely Excised Basal Cell Carcinoma of the Skin” (1). Our study evaluated the natural history of basal cell carcinomas that have been excised with a positive microscopic margin. In instances where follow-up is reliable, and the clinician is confident about the absence of residual disease, we recommend consideration of observing these lesions, as opposed to immediately treating them, since less than 50% of them relapse. When observed lesions do relapse, treatment at that time achieves the same local control as when lesions are treated immediately. The majority of patients with BCC are elderly, often encumbered with other illnesses, such that it would be advantageous to spare them potentially unnecessary treatments. Dr. Smee cautioned against observing incompletely excised BCC under three conditions: When the incomplete margin rests under a graft or flap. We specifically examined the issue of whether an increased local recurrence rate occurs when a graft or flap was performed, and contrary to our expectation, this was not found. Perineural invasion. This is an extremely rare feature in BCC’s and represents a very aggressive variant (2). Thus, when accompanied by a positive margin, we would anticipate that most clinicians would approach this lesion aggressively. BCC’s that occur at embryologic junctions have a notorious reputation for returning at depth. In this review, we did not study specifically whether lesions originated at an embryologic junction, but, given the difficulty in assessing presence of residual disease at this location, and their predilection for deep recurrences, most clinicians would approach such lesions aggressively. As emphasized in our original paper, this approach of observing incompletely excised BCC’s can only be made when one is confident that gross residual disease is not present, if the deep margin is not positive, and the lesion has not previously recurred, Dr. Smee has drawn attention to clinical situations whereby the physician may lack
1. J. Radiation
Oncology
0 Biology 0 Physics
could even be in a situation where we have less late effects at HDR than at LDR - for the same early effects. This phenomenon which, primafacie, appears contrary to basic radiobiological theory, depends on several factors, the main ones being the dose to the bladder/rectum relative to the treatment dose, and the relative curviness of the dose-response curves for early and late-responding tissues. This latter is usually described by “o/B” ratios; for the best available values of these parameters, we have shown that if the brachytherapy delivers less than about 75% of the treatment dose to the bladder/rectum, then if the HDR dose is reduced to give equal early effects to LDR, the HDR late effects should not be worse than the LDR late effects. Simply on these considerations, the smaller the number of fractions the better. In practice, however, other considerations, such as the need for tumor reoxygenation, would suggest use of no less than about five fractions. There is a third related factor to consider, which is that the short treatment time that is characteristic of HDR allows packing and retraction of the sensitive organs, estimated by Orton (4) to result in a 20% further decrease in rectal/bladder dose. This advantage results from an extra physical factor, on top of the biological factors discussed here. In light of these considerations, we would suggest that Orton’s suggestion (5) regarding the use of large numbers of fractions for HDR intracavitary brachytherapy of the cervix should be replaced by the following:
I. When the dose to the dose-limiting critical normal tissues (bladder/ rectum) is less than about 75% of the prescribed dose, for equal early response, HDR results in comparable (or less) late effects than LDR. 2. When HDR is indicated, about five fractions is probably optimal. There is no need for larger fraction numbers. 3. When HDR is indicated, HDR protocols comparable to those at LDR should be designed based on matching early rather than late effects. 4. For those patients where the dose to the bladder/rectum is comparable to the prescribed dose, HDR is contraindicated.
It should be noted, of course, that for other sites of interest for HDR (e.g., endometrial carcinoma), unless the dose-limiting organs receive a significantly smaller dose than the treatment dose, Orton’s comment is probably appropriate, and HDR might be expected to result in some loss in therapeutic efficacy. DAVIDJ. BRENNER,PH.D. ERIC J. HALL, D.Sc. Center for Radiological Research Columbia University 630 West 168th St. New York. NY 10032
1. Brenner, D. J.; Hall, E. J. Fractionated high dose-rate versus low dose-rate brachytherapy of the cervix. I. General considerations based on radiobiology. Brit. J. Radiol. 64:133-141; 1991. 2. Brenner, D. J.; Hall, E. J. Fractionated high dose-rate versus low dose-rate regimens for intracavitary brachytherapy of the cervix II. Equivalent regimes for combined brachytherapy and external irradiation. Int. J. Radiat. Oncol. Biol. Phys. (in press). 3. Dale, R. G. The use of small fraction numbers naecological afterloading: Some radiobiological J. Radiol. 63:290-294; 1990.
in high dose-rate gyconsiderations. Brit.
4. Orton, C. G. Remote Afterloading for cervix cancer: The physicist’s point of view. In: Martinez, A. A., Orton, C. G., Mould, R. F., eds. Brachytherapy HDR and LDR. The Netherlands: Nucletron; 1990. 5. Orton, C. G. HDR in gynecological applications: Dose and volume considerations. Int. J. Radiat. Oncol. Biol. Phys. 20:1379-1380; 1991. 6. Van Lancker, M.; Storme, G. Prediction of severe late complications in fractionated high dose rate brachytherapy in gynecological applications. Int. J. Radiat. Oncol. Biol. Phys. 20:1125-1129; 1990. 7. Withers, H. R.; Thames, H. D.; Peters, L. J. Differences in the fractionation response of acutely and late-responding tissues. In: Karcher,
Volume 22, Number
1. 1992
K. H., Kogelnik, H. D., Reinartz, G., eds. Progress in radio-oncology, vol. II. New York, NY: Raven Press; 1982:287-296.
FRACTIONATION IS IMPORTANT FOR HDR CERVIX CANCER BRACHYTHERAPY To the Editor: Drs. Brenner and Hall (2) make a very interesting observation concerning HDR brachytherapy when used for treatments when the dose to the limiting late-reacting normal tissues is below a certain “threshold” fraction of that to the tumor. Under these conditions, doses to normal tissues are so low, that increasing the number of HDR fractions no longer increases the therapeutic ratio. This is because cell survival curves for tumors and late-reacting normal tissues have different shapes (see their Fig. 1) and this leads to more sparing of tumor cells than normal cells by increasing fractionation. Up to this point, I completely agree with Drs. Brenner and Hall. But I completely disagree with them on the normal-tissue threshold dose below which this phenomenon occurs, and the subsequent relevance to cervix cancer radiotherapy. In their Letter to the Editor (2), they state that if the dose to rectal and bladder tissues is less than about 75% of the Point A dose, fractionation of the HDR treatments decreases the therapeutic ratio, that is, the threshold fraction is 0.75. They also state that Dale (4) agrees with them. However, they fail to report that, although Dale does agree that if the ratio of the rectal/bladder to Point A doses, which he calls the geometrical sparing factor m, is below a certain critical value, then fractionation decreases the therapeutic ratio; he shows graphically (Dale’s Fig. I) that this threshold value off is about 0.3, not 0.75. This is a vitally important difference, since average “hot-spot” rectal/bladder doses in HDR brachytherapy are typically 0.64.7 of the Point A doses (6). If Dale is correct, then increased fractionation is good, whereas if Brenner and Hall are right, it is not. Since both the Dale (4) and Brenner and Hall (2) studies both apply the linear-quadratic (L-Q) model to arrive at their conclusions, this difference is not due to the models they utilize but, as shown in the Appendix, is caused by the parameters (a/p values) used in these models. As shown in the Appendix, the critical threshold value off is given by:
where (u/B), and (o/B), are the a/B values for late-reacting normal tissues and tumor, respectively. Dale (4) states that he has used (a/B), = 3 Gy and (o/B), = IO Gy and this leads to a threshold value offof 0.3. Brenner and Hall do not state in their letter what a/B values they have assumed, but reference to an earlier paper (1) shows that they have probably used (n/B), = 4 Gy and (o/B), = 5.3 Gy, which leads to their threshold value offof 0.75. To return now to the original question: Does increased fractionation of HDR brachytherapy for cervix cancer improve the therapeutic ratio? According to the predictions of the L-Q model the answer is clearly yes, if one accepts the typical a/B values published for late-reacting tissues and tumors of 3 Gy and IO Gy, respectively, since their ratio is 0.3, which is well below the observed f values of 0.60.7 (6). However, since we really do not know these a/B values with any confidence, and since the L-Q model itself is probably only an approximation, the ultimate answer to this question has to come from clinical experience. Fortunately, we now have the benefit of the results of a recent survey of 56 institutions, treating over 17,000 Ca cervix patients with HDR brachytherapy (6), which shows that increased fractionation plays a very significant role in decreasing the risk of late complications. For example, for Pt. A doses 5 7 Gy/fraction the risk of severe late complications is observed to be 2-3 times lower than for doses > 7 Gyi fraction (6). In conclusion, therefore, I adhere to my previous contention that, for cervix cancer HDR treatments it is important to “keep the dose per,fraction down by giving manyfractions” (5). COLING. GRTON, PH.D. Gershenson Radiation Oncology Center Harper Hospital and Wayne State University 3990 John R. Street Detroit, MI 48201
223
Correspondence
APPENDIX Question:
for a fixed bioeffect dose to the tumor, does the bioeffect dose to normal late-reacting tissues increase or decrease with increased fractionation?
Answer:
the L-Q equations for bioeffect doses (ERDs) for tumor (subscript I) and late-reacting tissues (subscript I) are (3):
ERD, = Nd,(l + &)
(1)
ERD, = Nfd, (1 + &)
(2)
where d, is the dose/fraction to the tumor delivered in N fractions, and f is the geometrical sparing factor (4), that is, the ratio of the normal tissue to tumor doses. Substituting the value of N from eq( 1) and eq(2) yields the following equation for ERD, in terms of d,:
ERD, =
f (ERD,) (1 + A)-’
(1 + &)
Then the rate of change of ERD, with d, is obtained by differentiating ERD,, with respect to d,, realizing that ERD, is to be kept constant. This differentiation yields the equation:
AERD,)
WERD,) _
d(d,)-
Clearly,
pm, -
1. Where reconstruction of the surgically excised area is such as to place the histologically incomplete margin beneath transported skin, either as a skin graft or rotation flap. Recurrence is, thus, only detected at a time when it has grown through the overlying tissue. 2. Where the pathology report identifies the presence of perineural extension as a component of the incomplete margin. Although this is uncommon for BCC’s, it still occurs and, if untreated, can lead to quite significant problems later on. 3. Where the BCC abuts up to particular anatomical areas, such as the nasolabial groove, medial canthus, or post-auricular groove. In these areas, subsequent evidence of disease can, in fact, be quite bulky and distant to the initial site. This relates to the ability of BCC’s to extend along, but not destroy, tissue planes. The above comments are not meant to imply that all patients necessarily need to be treated, but the policy of observation with a histologically incomplete margin must be put in the appropriate clinical context. DR. ROBERT SMEE Department of Radiation Oncology Prince of Wales Hospital Randwick, NSW 2031
WP),l
(43, b/P), (1 +
ERD, increases
with increase
in d, (or decrease
A&B), - (a/B), is positive,
Hence, the threshold given by:
Dr. F. F. Liu et al. (1) They have endeavoured to establish a management approach to a frequently encountered clinical situation. In doing so, they are further refining the natural history of basal cell carcinomas (BCC’s) of the skin. However, I would view with concern the broad generalization of their article that observation can be adopted for all BCC’s incompletely excised. This policy can certainly apply in situations in which follow-up is available and adequate anatomical assessment is assured. There are, however, situations in which observation alone is clinically inappropriate. These circumstances are:
(cllp)I I
that is, f > cdpj
value off above which fractionation
f rhwshoid
1. Liu, F.-F.; Maki, E.; Warde, P.; Payne, D.; Fitzpatrick, P. A management approach to incompletely excised basal cell carcinoma of the skin. Int. J. Radiat. Oncol. Biol. Phys. 20:423-428; 1991.
in N) when
is desirable
is
_ (@)I (cl/P),
1 Brenner, D. J.; Hall, E. J. Fractionated high dose-rate versus low dose-rate brachytherapy of the cervix. I. General considerations based on radiobiology. Brit. J. Radiol. 64: 133-141; 1991. 2. Brenner, D. J.; Hall, E. J. HDR brachytherapy for carcinoma of the cervix: Fractionation considerations. Int. J. Radiat. Oncol. Biol. Physics. (in press). 3. Dale, R. G. The application of the linear-quadratic dose-effect equation to fractionated and protracted radiotherapy. Brit. J. Radiol. 58: 515-528; 1985. 4. Dale, R. G. The use of small fraction numbers naecological afterloading: Some radiobiological J. Radiol. 63:290-294; 1990.
in high dose-rate gyconsiderations. Brit.
5. Orton, C. G. HDR in gynecological applications: Dose and volume considerations. Int. J. Radiat. Oncol. Biol. Phys. 20:1379-1380; 1991. 6. Orton, C. G.; Seyedsadr, M.; Somnay, A. Comparison of high and low dose rate remote afterloading for cervix cancer and the importance of fractionation. Int. J. Radiat. Oncol. Biol. Physics 21:1425-1434; 1991.
CRITIQUE OF ‘A MANAGEMENT APPROACH TO INCOMPLETELY EXCISED BASAL CELL CARCINOMA OF THE SKIN’ To the Editor: I read with interest the recent article “A Management Approach to Incompletely Excised Basal Cell Carcinoma of the Skin” by
RESPONSE
TO DR. SMEE
To rhe Ediror: We thank Dr. R. Smee for his letter commenting on our paper “A Management Approach to Incompletely Excised Basal Cell Carcinoma of the Skin” (1). Our study evaluated the natural history of basal cell carcinomas that have been excised with a positive microscopic margin. In instances where follow-up is reliable, and the clinician is confident about the absence of residual disease, we recommend consideration of observing these lesions, as opposed to immediately treating them, since less than 50% of them relapse. When observed lesions do relapse, treatment at that time achieves the same local control as when lesions are treated immediately. The majority of patients with BCC are elderly, often encumbered with other illnesses, such that it would be advantageous to spare them potentially unnecessary treatments. Dr. Smee cautioned against observing incompletely excised BCC under three conditions: When the incomplete margin rests under a graft or flap. We specifically examined the issue of whether an increased local recurrence rate occurs when a graft or flap was performed, and contrary to our expectation, this was not found. Perineural invasion. This is an extremely rare feature in BCC’s and represents a very aggressive variant (2). Thus, when accompanied by a positive margin, we would anticipate that most clinicians would approach this lesion aggressively. BCC’s that occur at embryologic junctions have a notorious reputation for returning at depth. In this review, we did not study specifically whether lesions originated at an embryologic junction, but, given the difficulty in assessing presence of residual disease at this location, and their predilection for deep recurrences, most clinicians would approach such lesions aggressively. As emphasized in our original paper, this approach of observing incompletely excised BCC’s can only be made when one is confident that gross residual disease is not present, if the deep margin is not positive, and the lesion has not previously recurred, Dr. Smee has drawn attention to clinical situations whereby the physician may lack
