Intratumoral DNA Content Heterogeneity in Laryngeal Squamous Cell Carcinoma Adel K.
El-Naggar, MD; Virginia Lopez-Varela, MD;
Mario A. Luna, MD; Randal
samples obtained from sections cut by Michaels and Gregor's method obtained from 21 consecutive total laryngectomies for squamous cell carcinoma were studied for intratumoral DNA content heterogeneity or homogeneity. Concordant DNA ploidy was manifested in all samples of five (23.8%) carcinomas (two diploid and three aneuploid), while 16 carcinomas (76.2%) demonstrated a variable DNA ploidy (diploid and aneuploid). Analysis of cellular proliferative activity demonstrated remarkable intratumoral stability in both concordant and discordant carcinomas. These data indicate there is a considerable heterogeneity of DNA ploidy but the proliferative rate is relatively stable within the carcinomas. Clinical implications of our findings are also presented. (Arch Otolaryngol Head Neck Surg. 1992;118:169-173) \s=b\ Multiple tissue
C everal studies have suggested an important role for DNA content analysis in the biological assessment of malig¬ nant solid tumors.16 It has also been observed that the degree of such DNA abnormalities varies not only from one type of neoplasm to another2-7-8 but also among neoplasms of the
type.9"14 The latter observation may have led to conflict¬ on the utility of DNA analysis in certain ing types of solid neoplasms, including squamous cell carci¬ noma of the larynx. Few studies of DNA ploidy in laryngeal squamous cell car¬ cinomas have been published.15"17 In these articles, the mag¬ nitude of the DNA abnormalities as well as the relationship of these abnormalities to established clinicopathologic pa¬ same
assessments
rameters and
patient outcome have been inconclusive and often contradictory. In this study, we systematically exam¬
ined the extent of intratumoral DNA-content heterogeneity in laryngeal squamous cell carcinomas to seek possible clar¬ ification of the discrepancies in the published data. MATERIALS AND METHODS Twenty-one total laryngectomy specimens removed for squamous cell carcinoma and accessioned between 1987 and 1990 in the Department of Pathology, M. D. Ander-
Accepted
publication September 13, 1991. Departments of Pathology (Drs El-Naggar, Luna, and Batsakis) and Head and Neck Surgery (Dr Weber), The University for
From the
of Texas M. D. Anderson Cancer Center, Houston; and the Universidad Autonoma de Guadalajara (Mexico) (Dr Lopez-Varela). Presented in part at the 33rd annual meeting of the American Society for Head and Neck Surgery, Waikaloa, Hawaii, May 7-9,1991. Reprint requests to the Department of Pathology, Box 85, The University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Blvd, Houston, TX 77030 (Dr El-Naggar).
Weber, MD; John G. Batsakis,
MD
Cancer Center (Houston, Tex) formed the materials for this study. Tumor size, anatomic location, and re¬ gional extent were recorded. The specimens were then cut transversely according to the method of Michaels and Gregor (Figure).18 Demographic information was ob¬ tained from patient records. Staging of each carcinoma incorporated both clinical and pathologic components. A minimum of three and a maximum of five formalinfixed tissue samples from discrete different areas of a given neoplasm were excised, labeled according to loca¬ tion, and then embedded in paraffin. The carcinomas were histologically graded according to World Health son
Organization (Geneva, Switzerland) guidelines.19
FLOW CYTOMETRY from each primary carcinoma were sepa¬ samples for their DNA content. The tissues were then rately analyzed fixed in 4% formaldehyde for 2 to 8 hours and embedded in paraffin-paraplast with a melting point of 56°C. All blocks contained approximately 1.5x2.0-cm tissue. Nuclear sus¬ pensions of the neoplasms were prepared using a modified version of the method of Hedley et al20 and McLemore et al.21 In each sample of tissue selected for analysis, atleast75% was neoplasm; the remainder consisted of nonneoplastic ele¬ ments. In a given neoplasm, if possible, tissue blocks with more than 10% necrosis were avoided. Otherwise, the necrotic areas were mapped and nonnecrotic tissue was reembedded for flow cytometric analysis. Two 50-µ sections The
cut on a microtome (Reichert-Jung 2030, Cambridge In¬ struments GmbH, Heidelberg, Germany) and placed into
were
15 x 100-mm glass culture tubes (Corning [NY] Glass Works). The sections were deparaffinized by three 15-minute treatments in xylene analogue (Histo-Solve X,
Biochemical Sciences, Bridgeport, NJ). Solution volumes were from 3 to 5 mL per tube. Rehydration was performed with graded ethyl alcohol (twice in 100%, twice in 95%, once in 80%, and once in 70%) and then left in 50% ethyl alcohol overnight. Final hydration was achieved through two treat¬ ments with Dulbecco's phosphate-buffered saline (Gibco Laboratories, Grand Island, NY). The specimens were incu¬ bated for 30 minutes at 37°C in prewarmed 0.5% pepsin at a pH of 1.5. The enzymatic digestion was stopped with 100 µ of cold 250 mg/mL of pepstatin A (Sigma Chemical Co, St Louis, Mo). The tissue was disaggregated using 3-mL sy¬ ringes with 18-gauge needles and was filtered through 37-pm nylon-mesh filters (Small Parts Ine, Miami, Fla). The concentration of nuclei per milliliter was established using an automated cell counter (Coulter ZB1, Coulter Electronics, Hialeah, Fla) and dilutions made to obtain a 106-nuclei per mil-
Downloaded From: http://archotol.jamanetwork.com/ by a DALHOUSIE UNIVERSITY-DAL-11762 User on 06/20/2015
Hüter. Cytospin slides were prepared from each sample (unexposed to RNAse) and stained with Wright-Giemsa. The slides were reviewed to confirm the presence of at least 75% tumor nuclei as well as to establish nuclear integrity. A 900-pL aliquot of the final dilution was treated with 100 pL of 0.25 mg/mL of ribonuclease A (Worthington Biochemical, Freehold, NJ) for 30 minutes at 37CC. The tubes containing these cells were then iced and stained with 50 pL of 0.25% propidium iodide. Staining was performed at least 20 min¬ utes before analysis, and the entire run was completed
within 3 hours. Flow rates were adjusted to count approx¬ imately 75 nuclei per second. A peak vs integral bit map was employed to gate out doublets. In each sample, at least 10 000 nuclei were evaluated. The DNA histograms were analyzed using the boxogram gating procedure of Johnston et al.22 Cytologie software (Coulter, version 2.02) was used only in cases exhibiting excessive debris with no evidence of doublets. All tumors analyzed with the cytologie software used the same debris subtraction model. Nonneoplastic tissue was considered to constitute the biological diploid standard and the first G0/G1 popula¬ tion was used to denote the diploid stemline. The DNA index (DI) was defined as the ratio of the peak channel number of the test sample to the peak channel of the normal diploid control.23 By definition, the DI of a diploid population is 1.0. A tumor was considered aneu¬ ploid when a distinct second G0/G1 peak, accounting for at least 10% of the cells analyzed, was present (or any number of such separate peaks). In the two samples with multiple aneuploid stemlines (patient 7; Table 1) the DI of the dominant stemline was used. The proliferative frac¬ tion was determined as the percentage of cells in S-phase (after debris subtraction, when necessary) according to the method of Baisch et al.24 The coefficient of variation for the diploid populations ranged from 2.2 to 7.4 (mean,
4.2±1.7SD). Photomicrograph of a laryngectomy sequentially Michaels and Gregor's method.™ Table
1.—Clinicopathologic and
Flow
sectioned
Cytometrically
No. of
Patient/
Tumor
Age, y/
Size,
Sex
cm
Lymph
Stage
Blocks/Total Examined
by
given neoplasm.
Determined Features of
Histologie
Aneuploid
Node Metastasis
Tumor DNA heterogeneity was defined as the presence of both diploid and aneuploid stemlines and the presence of multiple abnormal stemlines in different parts of a
Grade
_,_
Diploid Aneuploid
1/33/F
2.5
Y
4/4
2/54/M
2.1
Y
2/4
M/P
3/85/M
4.0
Y
0/4
M
4/85/M
4.0
Y
4/5
M
5/75/M
4.0
2/4
M
6/71/M
7.0
Y
3/4
P
7/57/M
3.0
Y
4/4+
8/59/M
3.5
N
1/4
9/58/M
4.0
N
1/4
10/70/M
4.0
N
4/4
11/49/M
5.0
Y
3/4
12/67/M
3.0
13/33/M
4.0
14/63/F
3.0
Laryngeal Squamous Cell Carcinomas*
Mean
DNA
S-Phase
Index of
Aneuploid Diploid Aneuploid Components
_,_
All
P
2.3
P
1.48
1.5
4.3
4.3
W/M
6.4
M
6.0
5.0
Percentage of Abnormal Cells
(Aneuploid Component)
2.3
1.17
15.0
17.0
1.79
5.0
4.0
7.0
1.17
27.3
6.0
6.0
1.87
23.0
P
5.0
M
3.3
W
W
7.0
M
P
7.5
M/P
3.8
M
M
5.5
7.0
Y
3/4
M
M
5.0
Y
2/3
M
M
4.3
N
2/3
M
P
5.0
1.37
14.3
3.3
1.71
19.0
6.0
10.0
1.30
30.0
8.0
6.0
1.79
17.0
3.8
1.78
25.5
5.0
1.86
16.3
5.0
5.0
1.57
25.7
4.0
4.5
1.71
4.5
4.7
4.0
5.0
1.60
16.5
15/74/M
2.8
N
2/3
M
M
7.7
10.0
6.5
2.12
27.5
16/63/F
3.5
N
1/3
W
M
3.7
4.0
3.0
2.18
18.0
17/53/M
3.0
Y
2/3
M
P
5.7
11.0
3.0
1.37
22.5
18/69/M
5.0
Y
3/4
M
M
5.3
7.0
4.7
2.14
23.7
19/47/M
4.3
N
1/4
W
W
3.3
2.0
7.0
1.20
15.0
20/64/M
3.0
N
0/3
P
4.0
4.0
21/75/M
4.0
Y
2/4
M
8.0
8.0
1.22
21.5
*W indicates well differentiated; M, moderately differentiated; P, tTwo of the four blocks showed multiple stemlines.
M
8.0
poorly differentiated; Y, yes; and N,
no.
Downloaded From: http://archotol.jamanetwork.com/ by a DALHOUSIE UNIVERSITY-DAL-11762 User on 06/20/2015
Table 2.—Flow
Cytometric Characteristics of Laryngeal Squamous Cell Characteristic
Carcinoma Concordant and Discordant DNA
Range
Tested
Means, %
Mean ± SE
Ploidy
Category
of
S-phase fraction,
% Overall Concordant diploid (a) Mixed diploid and aneuploid
0.75 4.00 0.75 2.00 3.00 3.00
5.14±0.52 4.13±0.13 5.66±0.62 6.16±0.60 6.42 ±0.83 5.60 ±0.60
Diploid (b) Aneuploic (c) Concordant aneuploid (d)
to to to to to to
12.25 4.25 12.25 11.00 17.00 11.00
d b
c vs
d
.15 .005 .80 .3
a vs
a vs
b
vs c
DNA index 1.62 ±0.08 1.64 ±0.09 1.55±0.19
1.17 to 2.18 1.17 to 2.18 1.17 to 1.78
a vs
b
.72
Abnormal cells, % 19.33: 1.60 Overall* 19.23: 1.85 Mixed diploid and aneuploid (a) 19.83: 3.06 Concordant aneuploid (b) *Overall total samples of concordant and discordant samples.
4.50 to 30.0 1.50 to 30.0 15.00 to 25.5
a vs
b
.88
Overall* Mixed diploid and aneuploid (a) Concordant aneuploid (b)
Table 3.—Probability of Missing Aneuploidy per of in Number
Analyzed Samples
Squamous
Probability of Missing Aneuploidy,
No. of
per
Laryngeal
Cell Carcinoma
Samples Specimen 1
33
2
17
3
8
4
3
%
RESULTS
Clinicopathologic Findings Table 1 presents the clinicopathologic and flow cytometric characteristics of the laryngeal carcinomas. The patient population consisted of 18 (86%) men and three (14%) women who ranged in age from 33 to 85 years (mean, 62 years). The neoplasms ranged in size from 2.1
cm (mean, 3.7 cm; median, 3.5 cm) except for a 7.0-cm tumor. In general, men were older than women (66.6 vs 53.0 years) and they had a relatively larger tumor (3.8 vs 3.0 cm). The carcinomas had the following stage: 17 T3 and 4 T4. Twelve patients had metastatic lymph node involvement by carcinoma at the time of surgery. Fourteen (66.6%) neoplasms manifested a concordant histologie grade throughout. Seven (33.3%) displayed some degree of disparate histologie grade in different ar¬ eas of the neoplasm. Histologie grade did not correlate with DNA content either within or between tumors. However, only one evaluated area in each of the two well-differentiated carcinomas was aneuploid. Notably, all samplings of the two concordant diploid neoplasms were poorly or moderately differentiated.
to 5.0
DNA
Ploidy
Two (9.5%) neoplasms displayed a concordant diploid DNA content and three (14.3%) a concordant aneuploid DNA content in all samples analyzed. Sixteen neoplasms (76.2%) manifested diploid and aneuploid DNA content in different sampled areas. There was no difference between the Dis of different samples of the aneuploid neoplasms and only one neoplasm expressed more than one aneuploid stemline in two of four samples tested. The
percentage of cells with aneuploid DNA content ranged
from 4.5% to 30.0%, with a mean of 19.30% ± 1.6%. The percentage of cells in the proliferative phase ranged from 3.0% to 17.0%, with a mean of 5.5% (the mean for the diploid neoplasms was 4.1% ± 0.15% and for the aneuploid neoplasms it was 5.8% ± 0.6%).
Proliferative Rate Table 2 presents the relationship between proliferative rate characteristics of the DNA homogeneous and DNA heterogeneous laryngeal squamous cell carcinomas. Con¬ cordant diploid neoplasms manifested significantly lower S-phases than those of concordant aneuploid neoplasms. A significantly different S-phase percentage was noted between the concordant diploid neoplasms and the dip¬ loid component of the heterogeneous neoplasms. There were no significant differences between the S-phase of the diploid and the aneuploid DNA content samples in the heterogeneous group. There were also no statistical differences in S-phases between the concordantly aneu¬ ploid DNA neoplasms and the heterogeneous group. No statistical differences were observed between the mean Dis and the percentage of abnormal cells of the aneuploid DNA components of the heterogeneous tu¬ mors and those of the concordant group. Table 3 presents the probability of failing to obtain an aneuploid DNA content based on the number of samples analyzed. Thirty-three percent, 17%, 8.0%, and 3.0% of the specimens with aneuploid DNA content would have been missed had we analyzed one, two, three, and four tumor
samples, respectively.
COMMENT
heterogeneity, defined as the con¬ diploid and aneuploid DNA contents or the presence of multiple stemlines in aneuploid tumors in different regions, has been demonstrated in several solid neoplasms.6-2529 Accordingly, an aneuploid DNA Intratumoral DNA
current presence of
content may be missed in
a
substantial number of solid
neoplasms if only a single sample manifesting a diploid DNA is analyzed. The degree of DNA content abnormality and its portent in the clinical course of laryngeal squamous cell carcinoma has varied considerably in published articles.15"17 These
Downloaded From: http://archotol.jamanetwork.com/ by a DALHOUSIE UNIVERSITY-DAL-11762 User on 06/20/2015
differences can be attributed to intratumoral variation in the DNA content between tumor cells,9"14 and/or techni¬ cal and interpretive factors.30 To determine the clinical potenutility of DNA analysis in assessing the biological factors tid of laryngeal squamous cell carcinomas these have to be addressed. Our data indicate that a high percentage of laryngeal squamous cell carcinomas manifest an aneuploid DNA content. The percent is considerably higher than others have reported.1517 We believe that this is attributable to an underestimation of intratumoral DNA variability in ear¬ lier studies. Aneuploidy may be missed in 33% of the cases
if
only a single sample is analyzed. This possibility
four drops to 17.0%, 8.0%, and 3.0%, when two, three, orGold¬ tested. In a are of cancer, study laryngeal samples
smith et al,16 found discordant intratumoral DNA content in only two (8.6%) of 23 tumors. This may have been re¬ lated to the apparent high coefficients of variation in their data, as illustrated in their histograms. Low-degree ane¬ is too uploidy can be missed if the coefficient of variation Gold¬ wide.31 Moreover, it is not clear from the study by smith and coworkers how many samples were analyzed per tumor. The samples' spatial relationship was also not stated. Accuracy and reproducibility are vital for a valid anal¬ ysis of the flow cytometrically determined DNA content of a neoplasm. Intralaboratory and interlaboratory assays of duplicate samples have indicated a high degree of of others16 suggest reproducibility.32"34 Our data and those that differences in measurement or interpretive variables in laryngeal squamous cell carcinoma are not solely responsible for the differences in observed DNA ploidy between studies. Our data indicate a considerable regional homogeneity of the proliferative fraction in these neoplasms. This is evidenced by the agreement in percentages of cells in proliferative phase between the diploid and aneuploid components of heterogeneous tumors. Interestingly, however, the S-phase percentage in neoplasms with ho¬ mogeneous diploid DNA content differed significantly from those of homogeneously aneuploid and heteroge¬ neous DNA content tumors. These findings suggest that DNA ploidy may be less informative than proliferative activity in estimating the biological course and treatment response of these neoplasms.31
El-Naggar, MD; Virginia Lopez-Varela, MD;
Mario A. Luna, MD; Randal
samples obtained from sections cut by Michaels and Gregor's method obtained from 21 consecutive total laryngectomies for squamous cell carcinoma were studied for intratumoral DNA content heterogeneity or homogeneity. Concordant DNA ploidy was manifested in all samples of five (23.8%) carcinomas (two diploid and three aneuploid), while 16 carcinomas (76.2%) demonstrated a variable DNA ploidy (diploid and aneuploid). Analysis of cellular proliferative activity demonstrated remarkable intratumoral stability in both concordant and discordant carcinomas. These data indicate there is a considerable heterogeneity of DNA ploidy but the proliferative rate is relatively stable within the carcinomas. Clinical implications of our findings are also presented. (Arch Otolaryngol Head Neck Surg. 1992;118:169-173) \s=b\ Multiple tissue
C everal studies have suggested an important role for DNA content analysis in the biological assessment of malig¬ nant solid tumors.16 It has also been observed that the degree of such DNA abnormalities varies not only from one type of neoplasm to another2-7-8 but also among neoplasms of the
type.9"14 The latter observation may have led to conflict¬ on the utility of DNA analysis in certain ing types of solid neoplasms, including squamous cell carci¬ noma of the larynx. Few studies of DNA ploidy in laryngeal squamous cell car¬ cinomas have been published.15"17 In these articles, the mag¬ nitude of the DNA abnormalities as well as the relationship of these abnormalities to established clinicopathologic pa¬ same
assessments
rameters and
patient outcome have been inconclusive and often contradictory. In this study, we systematically exam¬
ined the extent of intratumoral DNA-content heterogeneity in laryngeal squamous cell carcinomas to seek possible clar¬ ification of the discrepancies in the published data. MATERIALS AND METHODS Twenty-one total laryngectomy specimens removed for squamous cell carcinoma and accessioned between 1987 and 1990 in the Department of Pathology, M. D. Ander-
Accepted
publication September 13, 1991. Departments of Pathology (Drs El-Naggar, Luna, and Batsakis) and Head and Neck Surgery (Dr Weber), The University for
From the
of Texas M. D. Anderson Cancer Center, Houston; and the Universidad Autonoma de Guadalajara (Mexico) (Dr Lopez-Varela). Presented in part at the 33rd annual meeting of the American Society for Head and Neck Surgery, Waikaloa, Hawaii, May 7-9,1991. Reprint requests to the Department of Pathology, Box 85, The University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Blvd, Houston, TX 77030 (Dr El-Naggar).
Weber, MD; John G. Batsakis,
MD
Cancer Center (Houston, Tex) formed the materials for this study. Tumor size, anatomic location, and re¬ gional extent were recorded. The specimens were then cut transversely according to the method of Michaels and Gregor (Figure).18 Demographic information was ob¬ tained from patient records. Staging of each carcinoma incorporated both clinical and pathologic components. A minimum of three and a maximum of five formalinfixed tissue samples from discrete different areas of a given neoplasm were excised, labeled according to loca¬ tion, and then embedded in paraffin. The carcinomas were histologically graded according to World Health son
Organization (Geneva, Switzerland) guidelines.19
FLOW CYTOMETRY from each primary carcinoma were sepa¬ samples for their DNA content. The tissues were then rately analyzed fixed in 4% formaldehyde for 2 to 8 hours and embedded in paraffin-paraplast with a melting point of 56°C. All blocks contained approximately 1.5x2.0-cm tissue. Nuclear sus¬ pensions of the neoplasms were prepared using a modified version of the method of Hedley et al20 and McLemore et al.21 In each sample of tissue selected for analysis, atleast75% was neoplasm; the remainder consisted of nonneoplastic ele¬ ments. In a given neoplasm, if possible, tissue blocks with more than 10% necrosis were avoided. Otherwise, the necrotic areas were mapped and nonnecrotic tissue was reembedded for flow cytometric analysis. Two 50-µ sections The
cut on a microtome (Reichert-Jung 2030, Cambridge In¬ struments GmbH, Heidelberg, Germany) and placed into
were
15 x 100-mm glass culture tubes (Corning [NY] Glass Works). The sections were deparaffinized by three 15-minute treatments in xylene analogue (Histo-Solve X,
Biochemical Sciences, Bridgeport, NJ). Solution volumes were from 3 to 5 mL per tube. Rehydration was performed with graded ethyl alcohol (twice in 100%, twice in 95%, once in 80%, and once in 70%) and then left in 50% ethyl alcohol overnight. Final hydration was achieved through two treat¬ ments with Dulbecco's phosphate-buffered saline (Gibco Laboratories, Grand Island, NY). The specimens were incu¬ bated for 30 minutes at 37°C in prewarmed 0.5% pepsin at a pH of 1.5. The enzymatic digestion was stopped with 100 µ of cold 250 mg/mL of pepstatin A (Sigma Chemical Co, St Louis, Mo). The tissue was disaggregated using 3-mL sy¬ ringes with 18-gauge needles and was filtered through 37-pm nylon-mesh filters (Small Parts Ine, Miami, Fla). The concentration of nuclei per milliliter was established using an automated cell counter (Coulter ZB1, Coulter Electronics, Hialeah, Fla) and dilutions made to obtain a 106-nuclei per mil-
Downloaded From: http://archotol.jamanetwork.com/ by a DALHOUSIE UNIVERSITY-DAL-11762 User on 06/20/2015
Hüter. Cytospin slides were prepared from each sample (unexposed to RNAse) and stained with Wright-Giemsa. The slides were reviewed to confirm the presence of at least 75% tumor nuclei as well as to establish nuclear integrity. A 900-pL aliquot of the final dilution was treated with 100 pL of 0.25 mg/mL of ribonuclease A (Worthington Biochemical, Freehold, NJ) for 30 minutes at 37CC. The tubes containing these cells were then iced and stained with 50 pL of 0.25% propidium iodide. Staining was performed at least 20 min¬ utes before analysis, and the entire run was completed
within 3 hours. Flow rates were adjusted to count approx¬ imately 75 nuclei per second. A peak vs integral bit map was employed to gate out doublets. In each sample, at least 10 000 nuclei were evaluated. The DNA histograms were analyzed using the boxogram gating procedure of Johnston et al.22 Cytologie software (Coulter, version 2.02) was used only in cases exhibiting excessive debris with no evidence of doublets. All tumors analyzed with the cytologie software used the same debris subtraction model. Nonneoplastic tissue was considered to constitute the biological diploid standard and the first G0/G1 popula¬ tion was used to denote the diploid stemline. The DNA index (DI) was defined as the ratio of the peak channel number of the test sample to the peak channel of the normal diploid control.23 By definition, the DI of a diploid population is 1.0. A tumor was considered aneu¬ ploid when a distinct second G0/G1 peak, accounting for at least 10% of the cells analyzed, was present (or any number of such separate peaks). In the two samples with multiple aneuploid stemlines (patient 7; Table 1) the DI of the dominant stemline was used. The proliferative frac¬ tion was determined as the percentage of cells in S-phase (after debris subtraction, when necessary) according to the method of Baisch et al.24 The coefficient of variation for the diploid populations ranged from 2.2 to 7.4 (mean,
4.2±1.7SD). Photomicrograph of a laryngectomy sequentially Michaels and Gregor's method.™ Table
1.—Clinicopathologic and
Flow
sectioned
Cytometrically
No. of
Patient/
Tumor
Age, y/
Size,
Sex
cm
Lymph
Stage
Blocks/Total Examined
by
given neoplasm.
Determined Features of
Histologie
Aneuploid
Node Metastasis
Tumor DNA heterogeneity was defined as the presence of both diploid and aneuploid stemlines and the presence of multiple abnormal stemlines in different parts of a
Grade
_,_
Diploid Aneuploid
1/33/F
2.5
Y
4/4
2/54/M
2.1
Y
2/4
M/P
3/85/M
4.0
Y
0/4
M
4/85/M
4.0
Y
4/5
M
5/75/M
4.0
2/4
M
6/71/M
7.0
Y
3/4
P
7/57/M
3.0
Y
4/4+
8/59/M
3.5
N
1/4
9/58/M
4.0
N
1/4
10/70/M
4.0
N
4/4
11/49/M
5.0
Y
3/4
12/67/M
3.0
13/33/M
4.0
14/63/F
3.0
Laryngeal Squamous Cell Carcinomas*
Mean
DNA
S-Phase
Index of
Aneuploid Diploid Aneuploid Components
_,_
All
P
2.3
P
1.48
1.5
4.3
4.3
W/M
6.4
M
6.0
5.0
Percentage of Abnormal Cells
(Aneuploid Component)
2.3
1.17
15.0
17.0
1.79
5.0
4.0
7.0
1.17
27.3
6.0
6.0
1.87
23.0
P
5.0
M
3.3
W
W
7.0
M
P
7.5
M/P
3.8
M
M
5.5
7.0
Y
3/4
M
M
5.0
Y
2/3
M
M
4.3
N
2/3
M
P
5.0
1.37
14.3
3.3
1.71
19.0
6.0
10.0
1.30
30.0
8.0
6.0
1.79
17.0
3.8
1.78
25.5
5.0
1.86
16.3
5.0
5.0
1.57
25.7
4.0
4.5
1.71
4.5
4.7
4.0
5.0
1.60
16.5
15/74/M
2.8
N
2/3
M
M
7.7
10.0
6.5
2.12
27.5
16/63/F
3.5
N
1/3
W
M
3.7
4.0
3.0
2.18
18.0
17/53/M
3.0
Y
2/3
M
P
5.7
11.0
3.0
1.37
22.5
18/69/M
5.0
Y
3/4
M
M
5.3
7.0
4.7
2.14
23.7
19/47/M
4.3
N
1/4
W
W
3.3
2.0
7.0
1.20
15.0
20/64/M
3.0
N
0/3
P
4.0
4.0
21/75/M
4.0
Y
2/4
M
8.0
8.0
1.22
21.5
*W indicates well differentiated; M, moderately differentiated; P, tTwo of the four blocks showed multiple stemlines.
M
8.0
poorly differentiated; Y, yes; and N,
no.
Downloaded From: http://archotol.jamanetwork.com/ by a DALHOUSIE UNIVERSITY-DAL-11762 User on 06/20/2015
Table 2.—Flow
Cytometric Characteristics of Laryngeal Squamous Cell Characteristic
Carcinoma Concordant and Discordant DNA
Range
Tested
Means, %
Mean ± SE
Ploidy
Category
of
S-phase fraction,
% Overall Concordant diploid (a) Mixed diploid and aneuploid
0.75 4.00 0.75 2.00 3.00 3.00
5.14±0.52 4.13±0.13 5.66±0.62 6.16±0.60 6.42 ±0.83 5.60 ±0.60
Diploid (b) Aneuploic (c) Concordant aneuploid (d)
to to to to to to
12.25 4.25 12.25 11.00 17.00 11.00
d b
c vs
d
.15 .005 .80 .3
a vs
a vs
b
vs c
DNA index 1.62 ±0.08 1.64 ±0.09 1.55±0.19
1.17 to 2.18 1.17 to 2.18 1.17 to 1.78
a vs
b
.72
Abnormal cells, % 19.33: 1.60 Overall* 19.23: 1.85 Mixed diploid and aneuploid (a) 19.83: 3.06 Concordant aneuploid (b) *Overall total samples of concordant and discordant samples.
4.50 to 30.0 1.50 to 30.0 15.00 to 25.5
a vs
b
.88
Overall* Mixed diploid and aneuploid (a) Concordant aneuploid (b)
Table 3.—Probability of Missing Aneuploidy per of in Number
Analyzed Samples
Squamous
Probability of Missing Aneuploidy,
No. of
per
Laryngeal
Cell Carcinoma
Samples Specimen 1
33
2
17
3
8
4
3
%
RESULTS
Clinicopathologic Findings Table 1 presents the clinicopathologic and flow cytometric characteristics of the laryngeal carcinomas. The patient population consisted of 18 (86%) men and three (14%) women who ranged in age from 33 to 85 years (mean, 62 years). The neoplasms ranged in size from 2.1
cm (mean, 3.7 cm; median, 3.5 cm) except for a 7.0-cm tumor. In general, men were older than women (66.6 vs 53.0 years) and they had a relatively larger tumor (3.8 vs 3.0 cm). The carcinomas had the following stage: 17 T3 and 4 T4. Twelve patients had metastatic lymph node involvement by carcinoma at the time of surgery. Fourteen (66.6%) neoplasms manifested a concordant histologie grade throughout. Seven (33.3%) displayed some degree of disparate histologie grade in different ar¬ eas of the neoplasm. Histologie grade did not correlate with DNA content either within or between tumors. However, only one evaluated area in each of the two well-differentiated carcinomas was aneuploid. Notably, all samplings of the two concordant diploid neoplasms were poorly or moderately differentiated.
to 5.0
DNA
Ploidy
Two (9.5%) neoplasms displayed a concordant diploid DNA content and three (14.3%) a concordant aneuploid DNA content in all samples analyzed. Sixteen neoplasms (76.2%) manifested diploid and aneuploid DNA content in different sampled areas. There was no difference between the Dis of different samples of the aneuploid neoplasms and only one neoplasm expressed more than one aneuploid stemline in two of four samples tested. The
percentage of cells with aneuploid DNA content ranged
from 4.5% to 30.0%, with a mean of 19.30% ± 1.6%. The percentage of cells in the proliferative phase ranged from 3.0% to 17.0%, with a mean of 5.5% (the mean for the diploid neoplasms was 4.1% ± 0.15% and for the aneuploid neoplasms it was 5.8% ± 0.6%).
Proliferative Rate Table 2 presents the relationship between proliferative rate characteristics of the DNA homogeneous and DNA heterogeneous laryngeal squamous cell carcinomas. Con¬ cordant diploid neoplasms manifested significantly lower S-phases than those of concordant aneuploid neoplasms. A significantly different S-phase percentage was noted between the concordant diploid neoplasms and the dip¬ loid component of the heterogeneous neoplasms. There were no significant differences between the S-phase of the diploid and the aneuploid DNA content samples in the heterogeneous group. There were also no statistical differences in S-phases between the concordantly aneu¬ ploid DNA neoplasms and the heterogeneous group. No statistical differences were observed between the mean Dis and the percentage of abnormal cells of the aneuploid DNA components of the heterogeneous tu¬ mors and those of the concordant group. Table 3 presents the probability of failing to obtain an aneuploid DNA content based on the number of samples analyzed. Thirty-three percent, 17%, 8.0%, and 3.0% of the specimens with aneuploid DNA content would have been missed had we analyzed one, two, three, and four tumor
samples, respectively.
COMMENT
heterogeneity, defined as the con¬ diploid and aneuploid DNA contents or the presence of multiple stemlines in aneuploid tumors in different regions, has been demonstrated in several solid neoplasms.6-2529 Accordingly, an aneuploid DNA Intratumoral DNA
current presence of
content may be missed in
a
substantial number of solid
neoplasms if only a single sample manifesting a diploid DNA is analyzed. The degree of DNA content abnormality and its portent in the clinical course of laryngeal squamous cell carcinoma has varied considerably in published articles.15"17 These
Downloaded From: http://archotol.jamanetwork.com/ by a DALHOUSIE UNIVERSITY-DAL-11762 User on 06/20/2015
differences can be attributed to intratumoral variation in the DNA content between tumor cells,9"14 and/or techni¬ cal and interpretive factors.30 To determine the clinical potenutility of DNA analysis in assessing the biological factors tid of laryngeal squamous cell carcinomas these have to be addressed. Our data indicate that a high percentage of laryngeal squamous cell carcinomas manifest an aneuploid DNA content. The percent is considerably higher than others have reported.1517 We believe that this is attributable to an underestimation of intratumoral DNA variability in ear¬ lier studies. Aneuploidy may be missed in 33% of the cases
if
only a single sample is analyzed. This possibility
four drops to 17.0%, 8.0%, and 3.0%, when two, three, orGold¬ tested. In a are of cancer, study laryngeal samples
smith et al,16 found discordant intratumoral DNA content in only two (8.6%) of 23 tumors. This may have been re¬ lated to the apparent high coefficients of variation in their data, as illustrated in their histograms. Low-degree ane¬ is too uploidy can be missed if the coefficient of variation Gold¬ wide.31 Moreover, it is not clear from the study by smith and coworkers how many samples were analyzed per tumor. The samples' spatial relationship was also not stated. Accuracy and reproducibility are vital for a valid anal¬ ysis of the flow cytometrically determined DNA content of a neoplasm. Intralaboratory and interlaboratory assays of duplicate samples have indicated a high degree of of others16 suggest reproducibility.32"34 Our data and those that differences in measurement or interpretive variables in laryngeal squamous cell carcinoma are not solely responsible for the differences in observed DNA ploidy between studies. Our data indicate a considerable regional homogeneity of the proliferative fraction in these neoplasms. This is evidenced by the agreement in percentages of cells in proliferative phase between the diploid and aneuploid components of heterogeneous tumors. Interestingly, however, the S-phase percentage in neoplasms with ho¬ mogeneous diploid DNA content differed significantly from those of homogeneously aneuploid and heteroge¬ neous DNA content tumors. These findings suggest that DNA ploidy may be less informative than proliferative activity in estimating the biological course and treatment response of these neoplasms.31
