- Email: [email protected]
Assessment of mitotic rate reporting in melanoma
The American Journal of Surgery (2012) 204, 969 –975
The Southwestern Surgical Congress
Assessment of mitotic rate reporting in melanoma Alison L. Burton, B.A., B.S.a, Michael E. Egger, M.D.a, Juliana E. Gilbert, B.S.a, Arnold J. Stromberg, Ph.D.b, Lee Hagendoorn, M.B.A.c, Robert C.G. Martin, M.D., Ph.D.a, Charles R. Scoggins, M.D., M.B.A.a, Kelly M. McMasters, M.D., Ph.D.a, Glenda G. Callender, M.D.a,* a
Division of Surgical Oncology, Department of Surgery, University of Louisville, 315 E. Broadway, Suite 312, Louisville, KY 40202, USA; bDepartment of Statistics, University of Kentucky, Lexington, KY, USA; cAdvertek, Inc, Louisville, KY, USA KEYWORDS: Melanoma; Mitotic rate; Prognosis
Abstract BACKGROUND: In patients with cutaneous melanoma, mitotic rate (MR) historically has been reported as the number of mitoses per high-power field (hpf) or per 10 hpf. The most recent revision of the American Joint Committee on Cancer melanoma staging system now incorporates MR and specifies that MR should be reported as mitoses per mm2, with a conversion factor of 1 mm2 equaling 4 hpf. However, because many pathologists continue to report MR in hpf units, we sought to compare the 2 conventions for reporting MR; this is important now that MR is used for staging and prognostic information. METHODS: A retrospective analysis was performed of a database that combined patients from a large multicenter study and our single-institution melanoma database. All patients with pathology reports that included MR were included. For patients with MR reported in hpf units, MR was converted to mitoses per 10 hpf. Statistical analysis was performed to test differences in Breslow thickness (BT), ulceration, sentinel lymph node (SLN) status, and overall survival (OS) (log-rank test) between the mitoses per mm2 group versus the mitoses per 10-hpf group. RESULTS: A total of 1,148 patients were identified; of these, 759 were reported as per mm2 and 389 were reported in hpf units. When patients were subdivided into categories of MR of 0, 1, or more than 1, there was no statistically significant difference in mean or median BT, ulceration, or SLN positivity within categories between patients with MR per mm2 versus patients with MR reported per 10 hpf. There was also no difference in OS between groups. Subdividing into smaller categories of MR of 0, 1, 2, 3, 4, 5, or more than 5 did not yield different results. CONCLUSIONS: Although the American Joint Committee on Cancer staging system reports a conversion factor for MR of 1 mm2 equals 4 hpf, no clinically meaningful differences in predictors of prognosis (BT, ulceration, SLN positivity) or OS were seen between groups when a conversion factor of 1 mm2 equaling 10 hpf was used. Therefore, for practical purposes, MR reported per 10 hpf approximates MR per mm2. © 2012 Elsevier Inc. All rights reserved.
Presented at the Southwestern Surgical Congress Annual Meeting, March 25-28, 2012, Rancho Palos Verdes, CA. * Corresponding author. Tel.: ⫹1-502-629-6950; fax: ⫹1502-629-3183. E-mail address: [email protected] Manuscript received March 12, 2012; revised manuscript May 22, 2012
0002-9610/$ - see front matter © 2012 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.amjsurg.2012.05.021
The incidence and lifetime risk of melanoma is increasing rapidly. One in 51 people will be diagnosed with melanoma at some point in their lifetime.1 Melanoma has become the sixth most common cause of cancer in the United States; in 2012, approximately 76,250 people will be diagnosed with melanoma, and 9,180 people will die from the
970
The American Journal of Surgery, Vol 204, No 6, December 2012
disease.2 Prognosis and treatment recommendations for patients with primary cutaneous melanoma depend largely on a number of clinicopathologic factors. Breslow tumor thickness (BT), sentinel lymph node (SLN) status, and ulceration have long been established as the most influential predictive factors in patients with melanoma. Mitotic rate (MR) also recently was found to be a significant predictor of prognosis, especially in patients with thinner melanomas.3– 6 In fact, the most recent American Joint Committee on Cancer (AJCC) guidelines (7th edition, 2009) include mitotic rate in the T staging criteria of thinner melanomas.7 A lack of standardization exists for how mitotic rate is reported. The most accurate method of measurement of MR is number of mitoses per mm2. MR historically has been reported as the number of mitoses per high-power field (hpf) or per 10 hpf. The hpf measurement lacks consistency because the exact measurement is highly dependent on the microscope used and the individual using the microscope. The most recent revision of the AJCC melanoma staging system now incorporates MR and specifies that MR should be reported as mitoses per mm2. This is done by identifying an area of dermis with the most mitoses or an MR “hot spot”7; if no hot spot can be identified then a representative mitosis can be used and the count can be extended into surrounding microscope fields until 1 mm2 has been assessed. Calibration of the microscope generally is indicated before MR measurement, and a conversion factor of 1 mm2 corresponding to an area of 4 hpf at a 400⫻ magnification has been suggested for the majority of, but not all, microscopes.8 Because many pathologists continue to report MR in hpf units, we sought a methodology to compare the 2 conventions for reporting MR. This has become especially important now that MR is used for staging and prognostic information.
Methods This was a retrospective analysis performed on the combined data of patients in our single-institution melanoma database as well as those patients enrolled in the Sunbelt Melanoma Trial (SMT), a large multi-institutional, prospective, randomized trial that enrolled more than 3,600 patients between 1996 and 2003 in 79 centers throughout North America. The SMT received institutional review board approval at each participating institution. Our institution’s melanoma database includes patients with melanoma who have received treatment at the University of Louisville. MR was determined for these patients by pathologists from our region using standard methods; MR was recorded in either mm2 or hpf units. Patients enrolled in the SMT were ages 18 to 70 years with cutaneous melanomas of 1 mm or more Breslow thickness and no clinical evidence of involved lymph nodes or metastatic disease. All patients underwent SLN biopsy; those with tumor-positive SLN by standard histology and
immunohistochemistry underwent completion lymphadenectomy. SLN-positive patients were randomized to 1 of 2 distinct protocols to receive adjuvant interferon-alfa-2b therapy versus no adjuvant therapy. Neither of the randomized protocols showed a statistically significant difference in disease-free survival or overall survival. Thus, the entire SMT patient population can be considered a homogenous group despite some patients having received adjuvant therapy, and some patients not having received adjuvant therapy. The technical details of the SMT have been described previously.9 In the SMT, central pathology review was performed on the first 10 cases from each participating institution as well as for patients with a tumor-positive SLN. Otherwise, MR status was determined at each participating institution using standard methods and reported per mm2 or per hpf. Analysis was performed on all patients with pathology reports that included MR. For those patients with MR reported in x hpf units, MR was converted to mitoses per 10 hpf by multiplying by 10/x. Statistical analysis was performed to test differences in BT, ulceration, SLN status, and overall survival (OS; log-rank test) between the mitoses per mm2 group versus the mitoses per 10-hpf group. Analysis of variance was used to compare continuous variables and the chi-square test was used for comparing proportions.
Results A total of 1,148 patients with pathology reports including MR data were identified. For the entire group, the median age was 50 years, and 57.8% were male. The median follow-up time was 58 months. For the entire group, the median BT was 1.58 mm, 47.5% of patients had an extremity melanoma, 27.4% of patients had ulceration, and 20.3%
Table 1 Histologic subtype, T-stage, and N-stage for the overall group N (%) Histologic subtype Superficial spreading Nodular Acral lentiginous Desmoplastic Lentigo maligna Other T stage T1 T2 T3 T4 N stage N0 N1 N2 N3
787 395 (50.2) 229 (29.1) 20 (2.5) 10 (1.3) 10 (1.3) 123 (15.6) 1,133 141 (12.4) 606 (53.4) 279 (24.6) 107 (9.4) 810 653 (80.6) 112 (13.8) 42 (5.2) 3 (.4)
A.L. Burton et al. Table 2
Mitotic rate reporting in melanoma
971
Number of patients per MR category
MR reported per mm2 MR category 0 1 N, % 100, 13.2 259, 34.1 MR reported or converted to per 10 hpf MR category 0 1 N, % 52, 13.4 65, 16.7
2 128, 16.9
3 71, 9.4
4 49, 6.5
5 37, 4.9
6 27, 3.6
7 18, 2.4
8 16, 2.1
9 7, 2.9
10 13, 1.7
⬎ 10 34, 4.5
2
3 28, 7.2
4 15, 3.9
5 21, 5.4
6
7
8
9 6, 1.6
10 36, 9.3
⬎ 10 94, 24.2
46, 11.8
had a positive sentinel lymph node. Histologic subtype and AJCC T-stage and N-stage for the overall group are shown in Table 1. Overall, 759 (66%) patients had MR reported as per mm2 and 389 (34%) patients had MR reported in hpf units. Of these, 123 patients had MR reported as the number of mitoses per 1 hpf, 1 patient had MR reported as the number of mitoses per 2 hpf, 1 patient had MR reported as the number of mitoses per 3 hpf, 1 patient had MR reported as the number of mitoses per 4 hpf, 5 patients had MR reported as the number of mitoses per 5 hpf, 1 patient had MR reported as the number of mitoses per 7 hpf, 1 patient had MR reported as the number of mitoses per 8 hpf, and 1 patient had MR reported as the number of mitoses per 20 hpf. The remaining 255 patients had MR reported as the number of mitoses per 10 hpf. Table 2 summarizes the number of patients per MR category. After converting the measurements reported in hpf units to mitoses per 10 hpf, a total of 184 patients had BT of 1 mm or less; of these, 100 (54.4%) patients had MR of 0 and 84 (45.6%) patients had MR greater than 0. A total of 949 patients had BT greater than 1 mm; of these, 40 (4.2%) patients had MR of 0 and 909 (95.8%) patients had MR greater than 0. Fifteen patients did not have BT recorded. These subgroups did not depend on units of measurement, because the categories “MR of 0” or “MR greater than 0” are independent of the unit of measurement. When patients were subdivided into categories of MR of 0, 1, or more than 1, there was no statistically significant difference in mean or median BT, ulceration, or SLN positivity within categories between patients with MR reported per mm2 versus patients with MR reported as number per 10 hpf (Table 3). There were also no difference in OS or disease-free survival (DFS) between groups (Figs. 1 and 2).
9, 2.3
8, 2.1
9, 2.3
Subdividing into smaller categories of MR of 0, 1, 2, 3, 4, 5, or more than 5 did not yield different results (Table 4). In addition, when only patients with thin melanomas (BT ⱕ 1 mm) were included, there was no difference in rate of positive SLN, or in DFS or OS when the group reported in mitoses per mm2 was compared with the group reported in mitoses per hpf (data not shown).
Comments The prognostic value of MR remains a topic of interest to melanoma experts throughout the world. In a study from the Sydney Melanoma Unit, Azzola et al10 suggested that MR is a more powerful prognostic indicator than ulceration, second only to primary tumor thickness in melanoma patients. Another study performed by Sondak et al11 at the University of Michigan showed that MR and younger age are predictors of SLN positivity in melanoma patients. Although MR has an established role as a prognostic indicator in thin melanomas, its role in thicker melanomas remains controversial. Zettersten et al12 showed MR to be a significant predictor of OS in melanomas of 4 mm or greater on univariate analysis, but it was not an independent predictor of OS on multivariate analysis. In SMT data (melanomas ⱖ 1 mm), we previously found MR to be predictive of SLN status as well as ulceration, but not predictive of DFS or OS.13 The new AJCC staging guidelines specify that melanomas 1.0-mm thick or less with an MR of 1 or greater should be classified as T1b based on analysis of data showing that MR is associated with worse prognosis and greater risk of positive SLN in this subgroup of patients. However, the new
Table 3 Clinicopathologic variables by MR group using conversion of 1 mm2 equals 10 hpf comparing 0, 1, and more than 1 Categories MR: number per mm2
N Median BT (mean) Ulcerated, N (%) SLN positive, N (%)
MR: number per 10 hpf
P values*
0
1
⬎1
0
1
⬎1
0
1
⬎1
100 .50 (.72) 3 (3.7) 2 (7.7)
259 1.36 (1.70) 33 (13.2) 41 (18.4)
400 1.90 (2.50) 149 (37.8) 74 (20.7)
52 .78 (.96) 4 (7.7) 2 (7.4)
65 1.50 (1.74) 9 (13.9) 6 (9.8)
272 2.00 (2.80) 104 (40.0) 64 (27.2)
.0650 .3148 .9687
.7494 .8914 .1113
.0713 .5749 .0670
*Comparing group reported as number per mm2 with group reported as number per 10 hpf in each category (0, 1, ⬎1).
972
The American Journal of Surgery, Vol 204, No 6, December 2012 tion regarding MR in a straightforward manner. Although we expect that MR will be reported in the future in mitosis per mm2 units more routinely, this convention is not uniformly used presently. Therefore, we sought to find a rea-
Figure 1 (A) Overall survival Kaplan–Meier curves comparing patients with 0 mitoses per mm2 and patients with 0 mitoses per 10 hpf (P ⫽ .4837). (B) Overall survival Kaplan–Meier curves comparing patients with 1 mitosis per mm2 and patients with 1 mitosis per 10 hpf (P ⫽ .9662). (C) Overall survival Kaplan–Meier curves comparing patients with more than 1 mitosis per mm2 and patients with more than 1 mitosis per 10 hpf (P ⫽ .8230).
AJCC guidelines do not include MR as a stage criterion for melanomas greater than 1 mm. Because MR has been established as a part of the new AJCC T-staging criteria of thin melanomas, it is essential to be able to obtain informa-
Figure 2 (A) Disease-free survival Kaplan–Meier curves comparing patients with 0 mitoses per mm2 and patients with 0 mitoses per 10 hpf (P ⫽ .2500). (B) Disease-free survival Kaplan–Meier curves comparing patients with 1 mitosis per mm2 and patients with 1 mitosis per 10 hpf (P ⫽ .6623). (C) Disease-free survival Kaplan–Meier curves comparing patients with more than 1 mitosis per mm2 and patients with more than 1 mitosis per 10 hpf (P ⫽ .0900).
A.L. Burton et al.
Mitotic rate reporting in melanoma
973
Table 4 Clinicopathologic Variables by MR Group Using Conversion of 1 mm2 ⫽ 10 hpf. Comparing 0, 1, 2, 3, 4, 5 and ⬎ 5 Categories Mitotic rate Breslow thickness N, per mm2 Median (mean) N (per 10 hpf) Median (mean) P value Ulceration N, per mm2 N (%) N, per 10 hpf N (%) P value SLN positive N, per mm2 N (%) N, per 10 hpf N (%) P value
0 88 .5 (.7) 52 .8 (1.0) .0650 0 81 3 (3.7) 52 4 (7.7) .3148 0 26 2 (7.7) 27 2 (7.4) .9687
1 258 1.4 (1.7) 65 1.5 (1.7) .7494 1 250 33 (10.5) 65 9 (13.9) .8914 1 223 41 (18.4) 61 6 (9.8) .1113
2 126 1.1 (1.7) 46 1.3 (1.6) .5583 2 126 33 (26.2) 46 7 (15.2) .1316 2 113 22 (19.4) 46 10 (21.7) .8278
sonable conversion factor for MR when reported in number of mitoses per hpf to the preferred standardized method of measurement per mm2. Although the AJCC cancer staging system reports a conversion factor for MR of 1 mm2 equaling 4 hpf,8 we found that a conversion factor of 1 mm2 equaling 10 hpf also is adequate; no statistically significant difference in predictors of prognosis (BT, ulceration, SLN positivity) or OS was seen between groups using a conversion of 1 mm2 equaling 10 hpf. Therefore, for practical purposes, MR reported per 10 hpf approximates MR per mm2. One of the potential limitations of our study was that most patients who were included in our analysis had melanomas of 1 mm or greater, which is not the subgroup to which MR most applies. In addition, this study was a post hoc analysis; thus, the trial was not powered to answer this specific question. Because of this, some subgroups had small numbers, which could lead to a false-negative result. Another limitation of our study was that MR reporting was not standardized and central pathology review was not performed for all cases. MR was determined at each individual institution using that institution’s standard methods. The SMT protocol did not specify the method for measurement of MR, and we did not determine what method was used to measure MR for this study. However, because the hot spot method was recommended in the 2010 AJCC melanoma staging guidelines and the SMT accrued patients from 1996 to 2003, it is unlikely that the hot spot method was used on all patients enrolled in SMT. In addition, no standardized calibration of microscopes was performed. The heterogeneity in methods of MR reporting and in microscope calibration certainly could lead to considerable inaccuracy; however, this heterogeneity is reflective of the methods in widespread use throughout the time period during which the SMT accrued patients. We applied the rec-
3 71 1.9 (2.4) 28 1.7 (2.6) .7424 3 70 19 (27.1) 28 9 (32.1) .6206 3 61 7 (11.5) 27 5 (18.5) .3746
4 49 1.8 (2.2) 15 2.0 (2.3) .7301 4 48 16 (33.3) 15 5 (33.3) 1.000 4 44 6 (13.6) 13 6 (46.2) .0115
5 37 2.5 (3.7) 21 2.1 (2.7) .2533 5 36 20 (55.6) 20 7 (35.0) .1402 5 33 4 (12.1) 19 3 (15.8) .7090
⬎5 115 2.5 (3.1) 162 2.6 (3.2) .6988 ⬎5 114 61 (53.5) 151 76 (50.3) .6083 ⬎5 106 35 (33.0) 130 40 (30.8) .7120
ommended AJCC conversion rate between MR measured in hpf and MR measured in mm2, with the understanding that this would be inaccurate. However, this could be interpreted as a potential strength of the study because the purpose of this study was to identify a conversion factor between what actually takes place in widespread practice and the recommended procedure. One direction for future study is to have central pathology review of a set of slides using the hot spot method and recording and comparing the results with the original reported MR (whether reported per hpf or per mm2). Another interesting direction for future study would be to use a mitosis-specific marker, such as the phosphohistone H3 immunostain, which has been reported to improve the reliability of MR measurement in other tumor types.14 An additional method of improving the reliability of MR measurement would be to report MR per number of melanoma cells instead of per a certain tumor area. However, this method might be quite difficult for pathologists to enact in actual practice. In conclusion, although the AJCC cancer staging system reports a conversion factor for MR of 1 mm2 equaling 4 hpf, no clinically meaningful differences in predictors of prognosis (BT, ulceration, SLN positivity) or OS were seen between groups when a conversion factor of 1 mm2 equaling 10 hpf was used. Therefore, for practical purposes, MR reported per 10 hpf approximates MR per mm2.
Acknowledgment This study included data from an investigator-initiated clinical trial supported by a grant from Schering Oncology Biotech. All data management for this trial was performed at University of Louisville.
974
The American Journal of Surgery, Vol 204, No 6, December 2012
References 1. National Cancer Institute. Surveillance Epidemiology and End Results. Available from: http://seer.cancer.gov/statfacts/html/melan.html. Accessed: March 7, 2012. 2. National Cancer Institute melanoma factsheet. Available from: http:// www.cancer.gov/cancertopics/types/melanoma. Accessed: March 7, 2012. 3. Kesmodel SB, Karakousis GC, Botbyl JD, et al. Mitotic rate as a predictor of sentinel lymph node positivity in patients with thin melanomas. Ann Surg Oncol 2005;12:449 –58. 4. Gimotty PA, Elder DE, Fraker DL, et al. Identification of high-risk patients among those diagnosed with thin cutaneous melanomas. J Clin Oncol 2007;25:1129 –34. 5. Francken AB, Shaw HM, Thompson JF, et al. The prognostic importance of tumor mitotic rate confirmed in 1317 patients with primary cutaneous melanoma and long follow-up. Ann Surg Oncol 2004;11: 426 –33. 6. Busam KJ. The prognostic importance of tumor mitotic rate for patients with primary cutaneous melanoma. Ann Surg Oncol 2004;11: 360 –1. 7. Edge SB, Byrd DR, Compton CC, et al, eds. AJCC Cancer Staging Manual. 7th ed. New York: Springer; 2009. 8. Gershenwald JE, Soong SJ, Balch CM. TNM staging system for cutaneous melanoma. . .and beyond. Ann Surg Oncol 2010;17:1475–7. 9. McMasters KM, Reintgen DS, Ross MI, et al. Sentinel lymph node biopsy for melanoma: how many radioactive nodes should be removed? Ann Surg Oncol 2001;8:6. 10. Azzola MF, Shaw HM, Thompson JF, et al. Tumor mitotic rate is a more powerful prognostic indicator than ulceration in patients with primary cutaneous melanoma: an analysis of 3661 patients from a single center. Cancer 2003;97:1488 –98. 11. Sondak VK, Taylor JM, Sabel MS, et al. Mitotic rate and younger age are predictors of sentinel lymph node positivity: lessons learned from the generation of a probabilistic model. Ann Surg Oncol 2004;11:247–58. 12. Zettersten E, Sagebiel RW, Miller JR, et al. Prognostic factors in patients with thick cutaneous melanoma (⬎ 4 mm). Cancer 2002;94: 1049 –56. 13. Roach BA, Burton AL, Mays MP, et al. Does mitotic rate predict sentinel lymph node metastasis or survival in patients with intermediate and thick melanoma? Am J Surg 2010;200:759 – 64. 14. Ikenberg K, Pfaltz M, Rakozy C, et al. Immunohistochemical dual staining as an adjunct in assessment of mitotic activity in melanoma. J Cutan Pathol 2012;39:324 –30.
Discussion Dr Robert Andtbacka (Salt Lake City, UT): Mitotic rate often is determined by viewing the microscopic slides at 400⫻ magnification (typically a 10⫻ ocular and a 40⫻ objective magnification). The recommended approach to enumeration of mitoses is to first find the area in the vertical growth phase containing the most mitotic figures, the socalled hot spot. After counting the mitoses in the hot spot, the count is extended to adjacent fields until an area corresponding to 1 mm2 is assessed. If no hot spot can be found and mitoses are randomly scattered throughout the lesion, then several different randomly chosen areas should be counted, summed, and the average listed as the mitotic rate. In tumors where the invasive component is less than 1 mm
in area, an attempt may be made to extrapolate a rate per square millimeter. The premise of this article is that hpf mitotic rate can be converted to mm2 by dividing the hpf mitotic rate by 4. I find this assumption flawed and inappropriate since there is no indication in the manuscript that the individual microscopes used were calibrated. The conversion rates can be very broad, ranging from 6.7 to 2.3 hpf equaling 1 mm2. How did the authors account for this in their data set collected at multiple institutions and with multiple different microscopes? The patients were partially derived from the SMT database and covered a time period of 1996 to 2003 at 79 different centers. Did the authors determine if the hot spot method was used to report the mitotic rate for patients that were reported as mitotic rate per hpf at these centers? In the 7th edition AJCC staging system, mitotic rate of 1/mm2 or greater is incorporated in the T1 (ⱕ1 mm Breslow thickness) lesions. The authors appear to use less than 1 mm and 1 mm or greater as their cut-offs in analyzing the 1,148 patients. How do the authors please explain this discrepancy? In the 7th edition AJCC staging system a mitotic rate cut-off of less than 1/mm2 and 1 mm2 or more is used. In their analysis, the authors use mitotic rate cut-off of 0/mm2, 1/mm2, and more than 1/mm2, could you please explain how these non AJCC cut-off were chosen? Dr Alison L. Burton (Louisville, KY): Thank you very much for taking the time to read and review our manuscript so carefully. We very much appreciate your comments and suggestions for improvement. In response to your question about the mitotic rate, the mitotic rate was determined at each individual institution using standard methods. There was no specified method for measurement or MR, and no standardization for calibration of microscopes. This could certainly lead to considerable inaccuracy; however, the method of MR reporting in the Sunbelt Melanoma Trial is reflective of the methods in widespread use during the time period during which the SMT accrued patients. We applied the recommended AJCC conversion rate, with the understanding that this would be inaccurate; however, a basic premise of this paper is that there has historically been considerable inaccuracy in the reporting of MR in melanoma. Your second point regarding the hot spot method, we did not determine what method was utilized to delineate MR. Because the hot spot method was recommended in the 2010 AJCC melanoma staging guidelines and the SMT accrued patients from 1996 to 2003, it is unlikely that the hot spot method was utilized on all patients enrolled in the SMT. However, again, this lack of standardization is representative of the heterogeneity of methods in widespread use during the time period of accrual for the patients in our study. Again, thank you for pointing out this idea. We agree that the lack of standardization for reporting mitotic rate leads to considerable inaccuracy, and certainly this could lead to the finding of no differences in our
A.L. Burton et al.
Mitotic rate reporting in melanoma
paper. We have considered a future study involving pathology review of the original slides utilizing the hot spot method and comparing the results with the original mitotic rate (whether reported per hpf or per mm2). Regarding your question about the non-AJC cut-offs, we chose to look at groups of MR of 0, 1, and more than 1 in order
975 to have meaningful groups for comparison. There was no difference between groups regardless of which subsets of MR we examined. Again, thank you for your questions and we will make sure these points will be incorporated in our revised manuscript.
The Southwestern Surgical Congress
Assessment of mitotic rate reporting in melanoma Alison L. Burton, B.A., B.S.a, Michael E. Egger, M.D.a, Juliana E. Gilbert, B.S.a, Arnold J. Stromberg, Ph.D.b, Lee Hagendoorn, M.B.A.c, Robert C.G. Martin, M.D., Ph.D.a, Charles R. Scoggins, M.D., M.B.A.a, Kelly M. McMasters, M.D., Ph.D.a, Glenda G. Callender, M.D.a,* a
Division of Surgical Oncology, Department of Surgery, University of Louisville, 315 E. Broadway, Suite 312, Louisville, KY 40202, USA; bDepartment of Statistics, University of Kentucky, Lexington, KY, USA; cAdvertek, Inc, Louisville, KY, USA KEYWORDS: Melanoma; Mitotic rate; Prognosis
Abstract BACKGROUND: In patients with cutaneous melanoma, mitotic rate (MR) historically has been reported as the number of mitoses per high-power field (hpf) or per 10 hpf. The most recent revision of the American Joint Committee on Cancer melanoma staging system now incorporates MR and specifies that MR should be reported as mitoses per mm2, with a conversion factor of 1 mm2 equaling 4 hpf. However, because many pathologists continue to report MR in hpf units, we sought to compare the 2 conventions for reporting MR; this is important now that MR is used for staging and prognostic information. METHODS: A retrospective analysis was performed of a database that combined patients from a large multicenter study and our single-institution melanoma database. All patients with pathology reports that included MR were included. For patients with MR reported in hpf units, MR was converted to mitoses per 10 hpf. Statistical analysis was performed to test differences in Breslow thickness (BT), ulceration, sentinel lymph node (SLN) status, and overall survival (OS) (log-rank test) between the mitoses per mm2 group versus the mitoses per 10-hpf group. RESULTS: A total of 1,148 patients were identified; of these, 759 were reported as per mm2 and 389 were reported in hpf units. When patients were subdivided into categories of MR of 0, 1, or more than 1, there was no statistically significant difference in mean or median BT, ulceration, or SLN positivity within categories between patients with MR per mm2 versus patients with MR reported per 10 hpf. There was also no difference in OS between groups. Subdividing into smaller categories of MR of 0, 1, 2, 3, 4, 5, or more than 5 did not yield different results. CONCLUSIONS: Although the American Joint Committee on Cancer staging system reports a conversion factor for MR of 1 mm2 equals 4 hpf, no clinically meaningful differences in predictors of prognosis (BT, ulceration, SLN positivity) or OS were seen between groups when a conversion factor of 1 mm2 equaling 10 hpf was used. Therefore, for practical purposes, MR reported per 10 hpf approximates MR per mm2. © 2012 Elsevier Inc. All rights reserved.
Presented at the Southwestern Surgical Congress Annual Meeting, March 25-28, 2012, Rancho Palos Verdes, CA. * Corresponding author. Tel.: ⫹1-502-629-6950; fax: ⫹1502-629-3183. E-mail address: [email protected] Manuscript received March 12, 2012; revised manuscript May 22, 2012
0002-9610/$ - see front matter © 2012 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.amjsurg.2012.05.021
The incidence and lifetime risk of melanoma is increasing rapidly. One in 51 people will be diagnosed with melanoma at some point in their lifetime.1 Melanoma has become the sixth most common cause of cancer in the United States; in 2012, approximately 76,250 people will be diagnosed with melanoma, and 9,180 people will die from the
970
The American Journal of Surgery, Vol 204, No 6, December 2012
disease.2 Prognosis and treatment recommendations for patients with primary cutaneous melanoma depend largely on a number of clinicopathologic factors. Breslow tumor thickness (BT), sentinel lymph node (SLN) status, and ulceration have long been established as the most influential predictive factors in patients with melanoma. Mitotic rate (MR) also recently was found to be a significant predictor of prognosis, especially in patients with thinner melanomas.3– 6 In fact, the most recent American Joint Committee on Cancer (AJCC) guidelines (7th edition, 2009) include mitotic rate in the T staging criteria of thinner melanomas.7 A lack of standardization exists for how mitotic rate is reported. The most accurate method of measurement of MR is number of mitoses per mm2. MR historically has been reported as the number of mitoses per high-power field (hpf) or per 10 hpf. The hpf measurement lacks consistency because the exact measurement is highly dependent on the microscope used and the individual using the microscope. The most recent revision of the AJCC melanoma staging system now incorporates MR and specifies that MR should be reported as mitoses per mm2. This is done by identifying an area of dermis with the most mitoses or an MR “hot spot”7; if no hot spot can be identified then a representative mitosis can be used and the count can be extended into surrounding microscope fields until 1 mm2 has been assessed. Calibration of the microscope generally is indicated before MR measurement, and a conversion factor of 1 mm2 corresponding to an area of 4 hpf at a 400⫻ magnification has been suggested for the majority of, but not all, microscopes.8 Because many pathologists continue to report MR in hpf units, we sought a methodology to compare the 2 conventions for reporting MR. This has become especially important now that MR is used for staging and prognostic information.
Methods This was a retrospective analysis performed on the combined data of patients in our single-institution melanoma database as well as those patients enrolled in the Sunbelt Melanoma Trial (SMT), a large multi-institutional, prospective, randomized trial that enrolled more than 3,600 patients between 1996 and 2003 in 79 centers throughout North America. The SMT received institutional review board approval at each participating institution. Our institution’s melanoma database includes patients with melanoma who have received treatment at the University of Louisville. MR was determined for these patients by pathologists from our region using standard methods; MR was recorded in either mm2 or hpf units. Patients enrolled in the SMT were ages 18 to 70 years with cutaneous melanomas of 1 mm or more Breslow thickness and no clinical evidence of involved lymph nodes or metastatic disease. All patients underwent SLN biopsy; those with tumor-positive SLN by standard histology and
immunohistochemistry underwent completion lymphadenectomy. SLN-positive patients were randomized to 1 of 2 distinct protocols to receive adjuvant interferon-alfa-2b therapy versus no adjuvant therapy. Neither of the randomized protocols showed a statistically significant difference in disease-free survival or overall survival. Thus, the entire SMT patient population can be considered a homogenous group despite some patients having received adjuvant therapy, and some patients not having received adjuvant therapy. The technical details of the SMT have been described previously.9 In the SMT, central pathology review was performed on the first 10 cases from each participating institution as well as for patients with a tumor-positive SLN. Otherwise, MR status was determined at each participating institution using standard methods and reported per mm2 or per hpf. Analysis was performed on all patients with pathology reports that included MR. For those patients with MR reported in x hpf units, MR was converted to mitoses per 10 hpf by multiplying by 10/x. Statistical analysis was performed to test differences in BT, ulceration, SLN status, and overall survival (OS; log-rank test) between the mitoses per mm2 group versus the mitoses per 10-hpf group. Analysis of variance was used to compare continuous variables and the chi-square test was used for comparing proportions.
Results A total of 1,148 patients with pathology reports including MR data were identified. For the entire group, the median age was 50 years, and 57.8% were male. The median follow-up time was 58 months. For the entire group, the median BT was 1.58 mm, 47.5% of patients had an extremity melanoma, 27.4% of patients had ulceration, and 20.3%
Table 1 Histologic subtype, T-stage, and N-stage for the overall group N (%) Histologic subtype Superficial spreading Nodular Acral lentiginous Desmoplastic Lentigo maligna Other T stage T1 T2 T3 T4 N stage N0 N1 N2 N3
787 395 (50.2) 229 (29.1) 20 (2.5) 10 (1.3) 10 (1.3) 123 (15.6) 1,133 141 (12.4) 606 (53.4) 279 (24.6) 107 (9.4) 810 653 (80.6) 112 (13.8) 42 (5.2) 3 (.4)
A.L. Burton et al. Table 2
Mitotic rate reporting in melanoma
971
Number of patients per MR category
MR reported per mm2 MR category 0 1 N, % 100, 13.2 259, 34.1 MR reported or converted to per 10 hpf MR category 0 1 N, % 52, 13.4 65, 16.7
2 128, 16.9
3 71, 9.4
4 49, 6.5
5 37, 4.9
6 27, 3.6
7 18, 2.4
8 16, 2.1
9 7, 2.9
10 13, 1.7
⬎ 10 34, 4.5
2
3 28, 7.2
4 15, 3.9
5 21, 5.4
6
7
8
9 6, 1.6
10 36, 9.3
⬎ 10 94, 24.2
46, 11.8
had a positive sentinel lymph node. Histologic subtype and AJCC T-stage and N-stage for the overall group are shown in Table 1. Overall, 759 (66%) patients had MR reported as per mm2 and 389 (34%) patients had MR reported in hpf units. Of these, 123 patients had MR reported as the number of mitoses per 1 hpf, 1 patient had MR reported as the number of mitoses per 2 hpf, 1 patient had MR reported as the number of mitoses per 3 hpf, 1 patient had MR reported as the number of mitoses per 4 hpf, 5 patients had MR reported as the number of mitoses per 5 hpf, 1 patient had MR reported as the number of mitoses per 7 hpf, 1 patient had MR reported as the number of mitoses per 8 hpf, and 1 patient had MR reported as the number of mitoses per 20 hpf. The remaining 255 patients had MR reported as the number of mitoses per 10 hpf. Table 2 summarizes the number of patients per MR category. After converting the measurements reported in hpf units to mitoses per 10 hpf, a total of 184 patients had BT of 1 mm or less; of these, 100 (54.4%) patients had MR of 0 and 84 (45.6%) patients had MR greater than 0. A total of 949 patients had BT greater than 1 mm; of these, 40 (4.2%) patients had MR of 0 and 909 (95.8%) patients had MR greater than 0. Fifteen patients did not have BT recorded. These subgroups did not depend on units of measurement, because the categories “MR of 0” or “MR greater than 0” are independent of the unit of measurement. When patients were subdivided into categories of MR of 0, 1, or more than 1, there was no statistically significant difference in mean or median BT, ulceration, or SLN positivity within categories between patients with MR reported per mm2 versus patients with MR reported as number per 10 hpf (Table 3). There were also no difference in OS or disease-free survival (DFS) between groups (Figs. 1 and 2).
9, 2.3
8, 2.1
9, 2.3
Subdividing into smaller categories of MR of 0, 1, 2, 3, 4, 5, or more than 5 did not yield different results (Table 4). In addition, when only patients with thin melanomas (BT ⱕ 1 mm) were included, there was no difference in rate of positive SLN, or in DFS or OS when the group reported in mitoses per mm2 was compared with the group reported in mitoses per hpf (data not shown).
Comments The prognostic value of MR remains a topic of interest to melanoma experts throughout the world. In a study from the Sydney Melanoma Unit, Azzola et al10 suggested that MR is a more powerful prognostic indicator than ulceration, second only to primary tumor thickness in melanoma patients. Another study performed by Sondak et al11 at the University of Michigan showed that MR and younger age are predictors of SLN positivity in melanoma patients. Although MR has an established role as a prognostic indicator in thin melanomas, its role in thicker melanomas remains controversial. Zettersten et al12 showed MR to be a significant predictor of OS in melanomas of 4 mm or greater on univariate analysis, but it was not an independent predictor of OS on multivariate analysis. In SMT data (melanomas ⱖ 1 mm), we previously found MR to be predictive of SLN status as well as ulceration, but not predictive of DFS or OS.13 The new AJCC staging guidelines specify that melanomas 1.0-mm thick or less with an MR of 1 or greater should be classified as T1b based on analysis of data showing that MR is associated with worse prognosis and greater risk of positive SLN in this subgroup of patients. However, the new
Table 3 Clinicopathologic variables by MR group using conversion of 1 mm2 equals 10 hpf comparing 0, 1, and more than 1 Categories MR: number per mm2
N Median BT (mean) Ulcerated, N (%) SLN positive, N (%)
MR: number per 10 hpf
P values*
0
1
⬎1
0
1
⬎1
0
1
⬎1
100 .50 (.72) 3 (3.7) 2 (7.7)
259 1.36 (1.70) 33 (13.2) 41 (18.4)
400 1.90 (2.50) 149 (37.8) 74 (20.7)
52 .78 (.96) 4 (7.7) 2 (7.4)
65 1.50 (1.74) 9 (13.9) 6 (9.8)
272 2.00 (2.80) 104 (40.0) 64 (27.2)
.0650 .3148 .9687
.7494 .8914 .1113
.0713 .5749 .0670
*Comparing group reported as number per mm2 with group reported as number per 10 hpf in each category (0, 1, ⬎1).
972
The American Journal of Surgery, Vol 204, No 6, December 2012 tion regarding MR in a straightforward manner. Although we expect that MR will be reported in the future in mitosis per mm2 units more routinely, this convention is not uniformly used presently. Therefore, we sought to find a rea-
Figure 1 (A) Overall survival Kaplan–Meier curves comparing patients with 0 mitoses per mm2 and patients with 0 mitoses per 10 hpf (P ⫽ .4837). (B) Overall survival Kaplan–Meier curves comparing patients with 1 mitosis per mm2 and patients with 1 mitosis per 10 hpf (P ⫽ .9662). (C) Overall survival Kaplan–Meier curves comparing patients with more than 1 mitosis per mm2 and patients with more than 1 mitosis per 10 hpf (P ⫽ .8230).
AJCC guidelines do not include MR as a stage criterion for melanomas greater than 1 mm. Because MR has been established as a part of the new AJCC T-staging criteria of thin melanomas, it is essential to be able to obtain informa-
Figure 2 (A) Disease-free survival Kaplan–Meier curves comparing patients with 0 mitoses per mm2 and patients with 0 mitoses per 10 hpf (P ⫽ .2500). (B) Disease-free survival Kaplan–Meier curves comparing patients with 1 mitosis per mm2 and patients with 1 mitosis per 10 hpf (P ⫽ .6623). (C) Disease-free survival Kaplan–Meier curves comparing patients with more than 1 mitosis per mm2 and patients with more than 1 mitosis per 10 hpf (P ⫽ .0900).
A.L. Burton et al.
Mitotic rate reporting in melanoma
973
Table 4 Clinicopathologic Variables by MR Group Using Conversion of 1 mm2 ⫽ 10 hpf. Comparing 0, 1, 2, 3, 4, 5 and ⬎ 5 Categories Mitotic rate Breslow thickness N, per mm2 Median (mean) N (per 10 hpf) Median (mean) P value Ulceration N, per mm2 N (%) N, per 10 hpf N (%) P value SLN positive N, per mm2 N (%) N, per 10 hpf N (%) P value
0 88 .5 (.7) 52 .8 (1.0) .0650 0 81 3 (3.7) 52 4 (7.7) .3148 0 26 2 (7.7) 27 2 (7.4) .9687
1 258 1.4 (1.7) 65 1.5 (1.7) .7494 1 250 33 (10.5) 65 9 (13.9) .8914 1 223 41 (18.4) 61 6 (9.8) .1113
2 126 1.1 (1.7) 46 1.3 (1.6) .5583 2 126 33 (26.2) 46 7 (15.2) .1316 2 113 22 (19.4) 46 10 (21.7) .8278
sonable conversion factor for MR when reported in number of mitoses per hpf to the preferred standardized method of measurement per mm2. Although the AJCC cancer staging system reports a conversion factor for MR of 1 mm2 equaling 4 hpf,8 we found that a conversion factor of 1 mm2 equaling 10 hpf also is adequate; no statistically significant difference in predictors of prognosis (BT, ulceration, SLN positivity) or OS was seen between groups using a conversion of 1 mm2 equaling 10 hpf. Therefore, for practical purposes, MR reported per 10 hpf approximates MR per mm2. One of the potential limitations of our study was that most patients who were included in our analysis had melanomas of 1 mm or greater, which is not the subgroup to which MR most applies. In addition, this study was a post hoc analysis; thus, the trial was not powered to answer this specific question. Because of this, some subgroups had small numbers, which could lead to a false-negative result. Another limitation of our study was that MR reporting was not standardized and central pathology review was not performed for all cases. MR was determined at each individual institution using that institution’s standard methods. The SMT protocol did not specify the method for measurement of MR, and we did not determine what method was used to measure MR for this study. However, because the hot spot method was recommended in the 2010 AJCC melanoma staging guidelines and the SMT accrued patients from 1996 to 2003, it is unlikely that the hot spot method was used on all patients enrolled in SMT. In addition, no standardized calibration of microscopes was performed. The heterogeneity in methods of MR reporting and in microscope calibration certainly could lead to considerable inaccuracy; however, this heterogeneity is reflective of the methods in widespread use throughout the time period during which the SMT accrued patients. We applied the rec-
3 71 1.9 (2.4) 28 1.7 (2.6) .7424 3 70 19 (27.1) 28 9 (32.1) .6206 3 61 7 (11.5) 27 5 (18.5) .3746
4 49 1.8 (2.2) 15 2.0 (2.3) .7301 4 48 16 (33.3) 15 5 (33.3) 1.000 4 44 6 (13.6) 13 6 (46.2) .0115
5 37 2.5 (3.7) 21 2.1 (2.7) .2533 5 36 20 (55.6) 20 7 (35.0) .1402 5 33 4 (12.1) 19 3 (15.8) .7090
⬎5 115 2.5 (3.1) 162 2.6 (3.2) .6988 ⬎5 114 61 (53.5) 151 76 (50.3) .6083 ⬎5 106 35 (33.0) 130 40 (30.8) .7120
ommended AJCC conversion rate between MR measured in hpf and MR measured in mm2, with the understanding that this would be inaccurate. However, this could be interpreted as a potential strength of the study because the purpose of this study was to identify a conversion factor between what actually takes place in widespread practice and the recommended procedure. One direction for future study is to have central pathology review of a set of slides using the hot spot method and recording and comparing the results with the original reported MR (whether reported per hpf or per mm2). Another interesting direction for future study would be to use a mitosis-specific marker, such as the phosphohistone H3 immunostain, which has been reported to improve the reliability of MR measurement in other tumor types.14 An additional method of improving the reliability of MR measurement would be to report MR per number of melanoma cells instead of per a certain tumor area. However, this method might be quite difficult for pathologists to enact in actual practice. In conclusion, although the AJCC cancer staging system reports a conversion factor for MR of 1 mm2 equaling 4 hpf, no clinically meaningful differences in predictors of prognosis (BT, ulceration, SLN positivity) or OS were seen between groups when a conversion factor of 1 mm2 equaling 10 hpf was used. Therefore, for practical purposes, MR reported per 10 hpf approximates MR per mm2.
Acknowledgment This study included data from an investigator-initiated clinical trial supported by a grant from Schering Oncology Biotech. All data management for this trial was performed at University of Louisville.
974
The American Journal of Surgery, Vol 204, No 6, December 2012
References 1. National Cancer Institute. Surveillance Epidemiology and End Results. Available from: http://seer.cancer.gov/statfacts/html/melan.html. Accessed: March 7, 2012. 2. National Cancer Institute melanoma factsheet. Available from: http:// www.cancer.gov/cancertopics/types/melanoma. Accessed: March 7, 2012. 3. Kesmodel SB, Karakousis GC, Botbyl JD, et al. Mitotic rate as a predictor of sentinel lymph node positivity in patients with thin melanomas. Ann Surg Oncol 2005;12:449 –58. 4. Gimotty PA, Elder DE, Fraker DL, et al. Identification of high-risk patients among those diagnosed with thin cutaneous melanomas. J Clin Oncol 2007;25:1129 –34. 5. Francken AB, Shaw HM, Thompson JF, et al. The prognostic importance of tumor mitotic rate confirmed in 1317 patients with primary cutaneous melanoma and long follow-up. Ann Surg Oncol 2004;11: 426 –33. 6. Busam KJ. The prognostic importance of tumor mitotic rate for patients with primary cutaneous melanoma. Ann Surg Oncol 2004;11: 360 –1. 7. Edge SB, Byrd DR, Compton CC, et al, eds. AJCC Cancer Staging Manual. 7th ed. New York: Springer; 2009. 8. Gershenwald JE, Soong SJ, Balch CM. TNM staging system for cutaneous melanoma. . .and beyond. Ann Surg Oncol 2010;17:1475–7. 9. McMasters KM, Reintgen DS, Ross MI, et al. Sentinel lymph node biopsy for melanoma: how many radioactive nodes should be removed? Ann Surg Oncol 2001;8:6. 10. Azzola MF, Shaw HM, Thompson JF, et al. Tumor mitotic rate is a more powerful prognostic indicator than ulceration in patients with primary cutaneous melanoma: an analysis of 3661 patients from a single center. Cancer 2003;97:1488 –98. 11. Sondak VK, Taylor JM, Sabel MS, et al. Mitotic rate and younger age are predictors of sentinel lymph node positivity: lessons learned from the generation of a probabilistic model. Ann Surg Oncol 2004;11:247–58. 12. Zettersten E, Sagebiel RW, Miller JR, et al. Prognostic factors in patients with thick cutaneous melanoma (⬎ 4 mm). Cancer 2002;94: 1049 –56. 13. Roach BA, Burton AL, Mays MP, et al. Does mitotic rate predict sentinel lymph node metastasis or survival in patients with intermediate and thick melanoma? Am J Surg 2010;200:759 – 64. 14. Ikenberg K, Pfaltz M, Rakozy C, et al. Immunohistochemical dual staining as an adjunct in assessment of mitotic activity in melanoma. J Cutan Pathol 2012;39:324 –30.
Discussion Dr Robert Andtbacka (Salt Lake City, UT): Mitotic rate often is determined by viewing the microscopic slides at 400⫻ magnification (typically a 10⫻ ocular and a 40⫻ objective magnification). The recommended approach to enumeration of mitoses is to first find the area in the vertical growth phase containing the most mitotic figures, the socalled hot spot. After counting the mitoses in the hot spot, the count is extended to adjacent fields until an area corresponding to 1 mm2 is assessed. If no hot spot can be found and mitoses are randomly scattered throughout the lesion, then several different randomly chosen areas should be counted, summed, and the average listed as the mitotic rate. In tumors where the invasive component is less than 1 mm
in area, an attempt may be made to extrapolate a rate per square millimeter. The premise of this article is that hpf mitotic rate can be converted to mm2 by dividing the hpf mitotic rate by 4. I find this assumption flawed and inappropriate since there is no indication in the manuscript that the individual microscopes used were calibrated. The conversion rates can be very broad, ranging from 6.7 to 2.3 hpf equaling 1 mm2. How did the authors account for this in their data set collected at multiple institutions and with multiple different microscopes? The patients were partially derived from the SMT database and covered a time period of 1996 to 2003 at 79 different centers. Did the authors determine if the hot spot method was used to report the mitotic rate for patients that were reported as mitotic rate per hpf at these centers? In the 7th edition AJCC staging system, mitotic rate of 1/mm2 or greater is incorporated in the T1 (ⱕ1 mm Breslow thickness) lesions. The authors appear to use less than 1 mm and 1 mm or greater as their cut-offs in analyzing the 1,148 patients. How do the authors please explain this discrepancy? In the 7th edition AJCC staging system a mitotic rate cut-off of less than 1/mm2 and 1 mm2 or more is used. In their analysis, the authors use mitotic rate cut-off of 0/mm2, 1/mm2, and more than 1/mm2, could you please explain how these non AJCC cut-off were chosen? Dr Alison L. Burton (Louisville, KY): Thank you very much for taking the time to read and review our manuscript so carefully. We very much appreciate your comments and suggestions for improvement. In response to your question about the mitotic rate, the mitotic rate was determined at each individual institution using standard methods. There was no specified method for measurement or MR, and no standardization for calibration of microscopes. This could certainly lead to considerable inaccuracy; however, the method of MR reporting in the Sunbelt Melanoma Trial is reflective of the methods in widespread use during the time period during which the SMT accrued patients. We applied the recommended AJCC conversion rate, with the understanding that this would be inaccurate; however, a basic premise of this paper is that there has historically been considerable inaccuracy in the reporting of MR in melanoma. Your second point regarding the hot spot method, we did not determine what method was utilized to delineate MR. Because the hot spot method was recommended in the 2010 AJCC melanoma staging guidelines and the SMT accrued patients from 1996 to 2003, it is unlikely that the hot spot method was utilized on all patients enrolled in the SMT. However, again, this lack of standardization is representative of the heterogeneity of methods in widespread use during the time period of accrual for the patients in our study. Again, thank you for pointing out this idea. We agree that the lack of standardization for reporting mitotic rate leads to considerable inaccuracy, and certainly this could lead to the finding of no differences in our
A.L. Burton et al.
Mitotic rate reporting in melanoma
paper. We have considered a future study involving pathology review of the original slides utilizing the hot spot method and comparing the results with the original mitotic rate (whether reported per hpf or per mm2). Regarding your question about the non-AJC cut-offs, we chose to look at groups of MR of 0, 1, and more than 1 in order
975 to have meaningful groups for comparison. There was no difference between groups regardless of which subsets of MR we examined. Again, thank you for your questions and we will make sure these points will be incorporated in our revised manuscript.