clear cell adenosquamous carcinoma of the cervix · clear cell adenosquamous carcinoma of the...

1591

Clear Cell Adenosquamous Carcinoma of the Cervix An Aggressive Tumor Associated with Human Papillomavirus- 18

Hisaya Fujiwara, M.D., Ph.D.,* Michelle Follen Mitchell, M.D., M.S.,t Jocelyne Arseneau, M.D.,$ Richard. J. Hale, M.D.,§ and Thomas C. Wright, Jr., M.D.*

Background. It is well recognized that adenocarcinomas and adenosquamous carcinomas of the cervix are frequently associated with human papillomavirus (HPV) -16 or -18. However, few studies have investigated associations between histologic variants of these tumors and specific types of HPV.

Methods. Eleven cases of cervical adenosquamous carcinoma with an unusual histologic appearance were characterized using histochemical and immunohistochemical stains. Sections were tested for the presence of HPV DNA using the polymerase chain reaction (PCR) and type specific primers for HPV-16 and -18. Clinical outcome was determined from a chart review.

Results. All tumors were histologically characterized by the presence of sheets of cohesive cells with prominent cell borders and a vacuolated or clear cytoplasm containing large amounts of glycogen. All tumors had focal gland formation and stained positive with mucicarmine stain. Using PCR, HPV-18 DNA was identified in all cases. The youngest patient was 24 years old and the oldest 74 years (mean, 43 years). Eight (73%) of the 11 patients have developed recurrent disease with a mean follow-up until recurrence of 9.5 months (range, 3-22 months). Seven (64%) of the 11 patients have died of their cervical tumors. Of the five patients with Stage IB disease, three (60%) have died of their cervical tumors.

Conclusions. A subset of invasive cervical adenosquamous carcinoma associated with HPV-18 that has a

From the Department of *Pathology, College of I’hysicians and Surgeons of Columbia University, New York, New York; the Division of tGynecologic Onrology, M.D. Anderson Cancer Center, Houston, Texas; the Department of $Pathology, Royal Victoria Hospital, McGill University, Montreal, Quebec, Canada; and the Department of §Pa- thology, St. Mary’s Hospital, Manchester, England.

Address for reprints: Thomas C. Wright, Jr., M.D., Department of Pathology, Room 16-428, College of Physicians and Surgeons of Columbia University, 630 W. 168th St., New York, NY 10032.

Received March 2, 1995; revision received June 22, 1995; ac- cepted June 22,1995.

distinctive histologic appearance and an aggressive clinical course is described. The term “clear cell adenosquamous carcinoma” is proposed for this unique variant of invasive cervical carcinoma. Cancer 1995; 76:1591-600.

Key words: cervical adenosquamous carcinoma, clear cell adenosquamous carcinoma, human papillomavirus type 18, histopathology.

Adenocarcinomas and adenosquamous carcinomas of the cervix have been the focus of a number of recent pathologic and epidemiologic studies. l,’ During the last several decades, the relative proportion of invasive adenocarcinoma to invasive squamous cell carcinoma of the cervix has been increasing3-’ Invasive adenocarcinomas of the cervix now account for approximately 10- 20% of all invasive cervical cancers. This apparent increase has been attributed to a number of factors including a disproportionate reduction in the number of invasive squamous cell carcinomas compared with adenocarcinomas secondary to widespread cytologic screening programs and to changes in the way in which cervical cancers are classified pathologically.1~2 How- ever, the absolute number of cervical adenocarcinomas has been reported to be increasing in the United King- dom,6 and there has been a significant increase in the incidence of invasive adenocarcinoma of the cervix in women younger than the age of 35 years in the United States.7

Like their squamous counterparts, invasive adenocarcinomas and adenosquamous carcinomas of the cervix frequently are associated with specific types of human papillomavirus (HI’V). Recent studies that have used either Southern blot or polymerase chain reaction (PCR) analyses to detect and type HPV have reported that between 31-83% of invasive adenocarcinomas and

1592 CANCER November 2 , 2995, Volume 76, No. 9

adenosquamous carcinomas of the cervix are associated with HPV and that there is a particularly strong association between these tumors and HPV-18.8-15 Little is known, however, about associations between specific types of HPV and specific histolop subtypes of cervical adenocarcinoma or adenosquamous carcinoma.

As part of an ongoing study assessing prognostic factors for invasive adenocarcinomas, we identified a group of tumors that originally were classified as adenocarcinomas of the cervix but which appear to be a subset of adenosquamous carcinoma with a distinctive histologic appearance. These tumors have not been previously described and are characterized by cells with prominent vacuolated cytoplasm containing large amounts of glycogen that form cohesive cellular sheets. All of these tumors are associated with HPV-18, and most had an aggressive clinical course. In this paper, we present the histologic features, clinical findings, and outcome of 11 patients with these tumors and propose that the term “clear cell adenosquamous carcinoma” should be used to refer to this unique variant of invasive cervical adenosquamous carcinoma.

Materials and Methods

Patients and Statistics

As part of a histopathologic study assessing the prognostic features of invasive adenocarcinomas of the cervix, we reviewed hematoxylin and eosin-stained histologic slides from 157 cases of invasive adenocarcinoma of the cervix. These cases were obtained from the pathology files of the College of Physicians and Surgeons of Columbia University (New York, NY), M.D. Ander- son Cancer Center (Houston TX), Royal Victoria Hospi- tal (Montreal, Quebec, Canada), and St. Mary’s Hospi- tal (Manchester, United Kingdom). All cases were diagnosed between 1983 and 1993. A variable number of slides was available for histopathologic review from each case. In some instances, the patient underwent a hysterectomy and the entire cervix was available for histopathologic assessment; in other cases, the patient was treated with radiation, and only a single cervical biopsy obtained before therapy was available for review. This study was approved by the Columbia Uni- versity Institutional Review Board.

Charts were reviewed for race, age, date of diagnosis, clinical International Federation of Gynecology and Obstetrics (FIGO) stage,” clinical lymph node status, lesion size, hemoglobin level at diagnosis, treatment, cell type, grade, last date of follow-up, survival status, and recurrence using a prepared clinical data abstract form. A literature search between the years of 2974 and

1994 was performed using MEDLINE to document that this entity was not previously reported.

Histopathology and Immunohistochemical Analysis

Formalin fixed, paraffin embedded tissues were used for histologic examination. Four-micron paraffin sections were stained with either hematoxylin and eosin, periodic acid-Schiff stain with (periodic acid-Schiff -d) and without (periodic acid-Schiff) digestion with diastase, or with mucicarmine stain. For immunohistochemistry, 4-pm sections were mounted on silane-coated slides (Digene Diagnostics, Silver Spring, MD) and were stained by the streptavidin-biotin method using primary monoclonal antibodies against cytokeratin (AE1/ AE3, mouse, 1:300, Boehringer Mannheim, Indianapo- lis, IN), vimentin (V9, mouse, 1:30, DAKO, Carpinteria, CA), carcinoembryonic antigen (CEA) (SP-651, mouse 1:200, Biogenex, San Ramon, CA), estrogen receptor (ERID5, mouse, 1:40, AMAC, Westrbrook, MA), progesterone receptor (PRI, mouse 1:1, CAS, Elmhurst, IL), and c-erbB-2 (monoclonal antibody 1, mouse, l : l O , Tri- ton, Alameda, CA). l7 Before staining, sections were treated with 0.3% hydrogen peroxide to block endoge- nous peroxidase activity and were incubated with normal goat serum. Sections used for identifying estrogen and progesterone receptors were treated with micro- wave irradiation before staining. The sections then were incubated for 18 hours at 4OC with the primary antibodies followed by treatment with biotinylated goat antimouse secondary antibody and streptavidin-biotin peroxidase complex. Finally, diaminobenzidine was ap- plied as chromogen to develop the peroxidase reaction. With each staining reaction, a positive control consist- ing of tissue known to contain the relevant antigen and a negative control of normal mouse immunoglobulin G, rather than the primary antibody, was included.

Polymerase Chain Reaction

Tissues were prepared for PCR by cutting three 10-pm sections from each formalin fixed, paraffin embedded tissue block, digesting the sections with proteinase K at 60° for 16 hours, and then separating the aqueous phase from the paraffin and undigested tissue by micro- centrifugation.” Aliquots of the aqueous phase were used for all PCR amplifications. Samples were analyzed for the presence of HPV-16 and -18 DNA using type- specific primers for the E6 and the L1 regions of these two HPV types. Two separate sets of type-specific primers were used. The sequences of the E6 region type-specific primer set were: HPV-16 positive strand 5’ ACC- GAAAACCGGTTAGTATAAAAGC 3‘; HPV- 16 negative strand 5’ ATAACTGTGGTAACTTTCTGGGTC 3‘;

Clear Cell Adenosquamous Carcinoma/Fujizuara et al. 1593

HPV-18 positive strand, 5’ CGGTCGGGACCGAA- AACGGTG 3’; and HPV-18 negative strand, 5’ CGT- GTTGGATCCTCAAAGCGCGCC 3’. The sequences of the L1 region type-specific primer set were HPV-16 positive strand 5’ TGCTAGTGCTTATGCAGCAA 3’; HPV-16 negative strand 5’ ATTTACTGCAACATTG- GTAC 3’; HPV-18 positive strand, 5‘ AAGGATGCT- GCACCGGCTGA 3’; and HPV-18 negative strand, 5’ CACGCACACGCTTGGCAGGT 3’”’ Ten pl of aqueous sample was added to 90 pl of reaction buffer containing a final concentration of 50 mM of potassium chloride, 4 mM of magnesium chloride, 10 mM of Tris, pH 8.5, 200 p M of dNTP, 0.2 pM of primers and 2.5 U of Taq DNA polymerase (Perkin-Elmer Cetus Corp., Foster City, CA). DNA was amplified for 35 cycles using the following parameters: denaturation for 1 minute at 94OC, reannealling for 1.5 minutes at 56OC, and extension for 1.5 minutes at 72OC for the first 34 cycles and extension for 3.0 minutes for Cycle 35. All PCR reac- tions contained appropriate positive and negative con- trols. Polymerase chain reaction products were electro- phoresed using an 8% native polyacrylamide gel and were stained with ethidium bromide. The 11 cases included in this series were part of a larger series of 157 tumors originally classified as cervical adenocarcinomas that we analyzed for HPV DNAusing this PCR method. Of the 157 cases, HPV-16 was detected in approximately one-third, HPV- 18 was detected in another third, and neither HPV-16 nor -18 was detected in the rest.

Results

HPV Typing and Histologic Features

The 11 cases of clear cell adenosquamous carcinoma described in this series were identified as part of a histopathologic review of 15 7 invasive cervical adenocarcinomas and, therefore, appeared to be relatively uncommon tumors. HPV-18 DNA was detected using PCR with both sets of type-specific primers in all 11 tumors (Fig. l), and all 11 had a similar histologic appearance by analysis of routine histologic sections.

The characteristic histologic pattern of clear cell adenosquamous carcinoma was that of a solid neoplasm composed of sheets of cohesive cells with prominent cell borders and a vacuolated or clear cytoplasm (Fig. 2a and 2b). The sheets of vacuolated cells frequently were subdivided by connective tissue septa, which in some cases contained a marked lymphocytic infiltrate. This often resulted in a lobular appearance (Fig. 3a). In some cases, large numbers of polymorpholeukocytes infiltrated the solid masses of tumor cells. In two cases, cells in the solid areas had abundant eosinophihc cytoplasm and were spindle-shaped,

W P 1 2 3 4 5 6 7 8 9 1 0 1 1 M

-21 7

-1 80

-147

-110

-90 -76 -67

Figure 1. Polyacrylamide gel (So/,) of polymerase chain reaction amplification products obtained using primers for the E6 region of human papillomavirus (HPV)-16 and -18. Lane W, water control; Lane P, HPV-18 plasmid DNA; Lanes 1-1 1, DNA extracted from cases 1-11; M: molecular weight markers.

suggesting an attempt at squamous differentiation. How- ever, in none of the cases was clear-cut hstologic evidence of squamous differentiation (e.g., keratin pearls, prominent intercellular bridges, or individual cell keratinization) identified. Highly atypical tumor cells with giant nuclei were focally present in most of the cases as were numer- ous mitoses (Fig. 3d). The nuclei of the tumor cells tended to be large and had considerable cell-to-cell variation in size and shape. Although prominent nucleoli occasionally were observed, these were not a characteristic feature.

All tumors initially were diagnosed as primary cervical adenocarcinomas. Focal gland formation was observed in all cases and was pronounced in 3 (27%) of the 11 (Fig. 3c). However, in three other cases, architectural evidence of glandular differentiation was quite minimal and only was detected by careful review (Fig. 3b). Papil- lary and tubulocystic areas typical of clear cell carcinoma were absent in all cases. In no areas did the cytoplasm assume the amphophilic appearance characteristic of glassy cell carcinoma, nor were hobnail cells, which typi- cally are associated with clear cell carcinomas, present. In none of the 11 cases was adjacent cervical intraepithelial neoplasia or adenocarcinoma in situ observed.

Histochemical and Immunohistochemical Findings

Sufficient tissue remained in the paraffin blocks after HPV analysis to allow histochemical and immunohistochemical studies in 7 of the 11 cases. These seven cases were analyzed using standard histochemical and immunohistochemical methods. The cytoplasm

1594 CANCER November 7,1995, Volume 76, No. 9

of the vacuolated cells contained large amounts of glycogen that stained positively with periodic acid- Schiff stain and was removed by digestion with diastase (Fig. 4a and 4b). Mucicarmine stains demonstrated intracellular mucin in all cases. In most, the mucicarmine staining was focal and weak, but in several, it was more pronounced (Fig. 4c and 4d). The immunohistochemical findings are summarized in Table 1 . In all cases, strong cytokeratin reactivity was present in the cytoplasm of the majority of the tumor cells (Fig. 5a). Carcinoembryonic antigen also was

Figure 2. Characteristic of clear cell adenosquamous carcinoma of the cervix (A, objective magnification X20; B, objective magnification X40).

demonstrated in all cases, but CEA staining was more focal than that observed for cytokeratin (Fig. 5b). In all seven cases, the tumor cells lacked vimentin. Al- though immunoreactivity for estrogen and progesterone receptors was detected using a streptavidin-biotin immunohistochemical method in cells within the stroma, the tumor cells were uniformly estrogen and progesterone receptor negative with this method. Similarly, no immunoreactivity for c-erbB-2 was detected in the seven tumors using a streptavidin-biotin immunohistochemical method (Table 1).

Figure 3. Histologic appearance of clear cell adenosquamous carcinoma of the cervix. Connective tissue septa (A) frequently subdivide the tumor and cause a lobular appearance. All tumors had gland formation (Band C). Highly atypical tumor cells (D) were present focally in many of the cases. (A, 8, and C, objective magnification X20; D, objective magnification X40).

Clear Cell Adenosquamous Carcinoma/Fujiwara et al. 1595

Figure 4. Periodic acid--Schiff stain either before (A) or after (B) diastase digestion. Intracellular mucin could be identified using the mucicarmine stain (C and D) (objective magnification X40).

Clinical Findings and Course

The clinical findings and course of the 11 patients are summarized in Table 2. The youngest patient was 24 years of age and the oldest was 74 years of age (mean age, 43 years). Five of the women were white, three were African American, and three were Hispanic. Five of the tumors were classified as clinical Stage IB, one as Stage IIA, two as Stage IIB, and two as Stage IIIB. One of the tumors was unstaged. Tumor size ranged from 3.0-8.0 cm (mean size, 5.7 cm) and lymph node metas- tases were found in 3 (27%) of the 11 cases. One of the five patients with Stage IB disease was treated with radical hysterectomy alone, and two were treated with

radiation therapy alone. The other two patients with Stage IB disease underwent radical hysterectomies that were followed by radiation therapy, and one was also treated with a course of chemotherapy. The six patients with higher stage or unstaged disease all were treated with radiation therapy either alone (n = 3) or a combi- nation of radical hysterectomy (n = 1) or chemotherapy (n = 2).

Eight (73%) of the 11 patients developed recurrent disease with a mean follow-up until recurrence of 9.5 months (range, 3-22 months). Seven (64%) of the 11 patients died of their cervical tumors. Survival in these seven ranged from 8-30 months (mean survival, 16.8 months). As of this writing, three (27%) of the patients

1596 CANCER November 1,1995, Volume 76, No. 9

Table 1. Histochemical and Immunohistochemical Characteristics of Clear Cell Adenosquamous Carcinoma

Estrogen Progesterone cerb Patient PAS PAS-d Mucicarmine Cytokeratin Vimentin CEA receptor receptor B-2

2 +++ ++ +++ - - +++ + 3 +++ +++ +++

5 ++ + +++ - - ++ ++ 6 +++ + +t + + 7 +++ +++ +++ + 9 t + +++

- - -

- - - - -

- - -

- - - - - - - - - -

- - - - -

i n +ti - + +++ - - - - + PAS: periodic acid Schiff stain; PAS-d: periodic acid Schiff stain after diastase digestion; CEA: carcinoembryonic antigen; -: no reactive cells; +: less than 1/3 of tumor cells are reactive; ++: 1/3-2/3 of tumor cells are reactive; +++: more than 2/3 of tumor cells are reactive.

were alive with no evidence of disease at a mean length of follow-up of 44.6 months, and one patient was alive with recurrent disease 4 months after therapy. Of the five patients with Stage IB disease, three (60%) died of their cervical tumors.

Discussion

These 11 carcinomas of the cervix appear to comprise a new histopathologic entity not previously reported, which we have termed clear cell adenosquamous carcinoma. In our opinion, the formation of a new histopathologic category for these tumors is justified because all 11 cases have a unique and distinctive histologic appearance, all are associated with HPV-18, and, most im- portantly, as a group, these tumors were associated with an aggressive clinical course.

The primary site for all 11 tumors was thought to be the cervix. This interpretation was based on the presence of a cervical tumor mass and the lack of endometrial involvement as determined from either a hysterectomy specimen obtained at the time of radlcal surgery or from endometrial curettings obtained at the time of surgical staging.” An additional finding supporting the endocervical origin, as opposed to an endometrial ori-

gin, was the lack of staining for the intermediate fila- ment vimentin and the focal positivity for CEA observed in all cases for which blocks were available for immunohistochemical analysis. Vimentin is detected in approximately 60% of invasive adenocarcinomas of the endometrium but is usually absent in primary endocervical adenocarcinomas.20,21 Conversely, staining for CEA is relatively uncommon in invasive adenocarcinomas of the endometrium but often is observed in primary endocervical adenocarc inoma~.~~,~~

All of the tumors in this series were architecturally poorly differentiated. Most contained regions composed of poorly differentiated cells lacking the clear or vacuolated cytoplasm characteristic of clear cell adenosquamous carcinoma. To be classified as a clear cell adenosquamous carcinoma, we required that at least 70% of the tumor be composed of sheets of cells with vacuolated or clear cytoplasm. We classified these tumors as adenosquamous carcinomas rather than undiff erenti- ated carcinomas for two reasons. First, all 11 cases had foci of glandular formation by analysis of routine hematoxylin and eoisin-stained sections. The recent Armed Forces Institute of Pathology (AFIP) and the World Health Organization’s (WHO) classification systems for invasive cervical cancer use the presence of

Figure 5. lmmunohistochemical stain for cytokeratin (A) or carcinoembryonic antigen (B) (objective magnification X40).

Clear Cell Adenosquamous Carcinoma/Fujiwara et aI. 1597

Table 2. Clinical Characteristics of Patients With Clear Cell Adenosquamous Carcinoma

Length of Lymph

Age FIGO Tumor nodal Time of I’atient (vrs) staee Race Theraov size (cm) status recurrence follow-up Status

1 37 I B White RT 5.0 Negative 9 mos. 14 mos. DOD 2 24 IB White OPE, RT 3.0 Negative 22 mos. 22 mos. DOD

3 44 I B Hispanic OPE, RT 8.0 Positive 9 mos. 30 mos. DOD 4 37 I B White RT NA Negative 59 mos. NED 5 41 IB White OPE 5.0 Negative 51 mos. NED 6 35 IIA African-American RT 8.0 Negative 24 mos. NED 7 29 IIB White OPE, RT 5.0 Positive 9 mos. 17 mos. DOD 8 66 IIB African-American RT, CHT NA Negative 10 mos. 16 mos. DOD 9 31 IIIB Hispanic RT, CHT 6.0 Positive 8 mos. 8 mos. DOD

10 55 1 IIB African-American RT NA Negative 6 mos. 11 mos. DOD 11 74 NA Hispanic RT NA Negative 3 mos. 4 mos. AWD

FIGO: International Federation of Gynecology and Obstetrics; RT: radiation therapy; DOD: dead of disease; A W D alive with disease; OPE: operation; CHT: chemotherapy; NA: not available; NED: no evidence of disease.

CHT

glandular elements as a criteria for adenocarcinomatous differentiati~n.~~,’~ In addition, in seven (100%) of the seven cases for which paraffin blocks were available, mucin was identified using a mucicarmine stain. Al- though there is considerable controversy as to the classification of cervical cancers that produce mucin but lack architectural evidence of glandular differentiation, mucin production generally is considered to be evidence of adenocarcinomatous differentiation in carcinomas at most clinical site^.^^,'^

The second reason for classifying these tumors as an adenosquamous carcinoma is that the solid sheets of vacuolated (e.g./ glycogenated) cells characteristic of these tumors are similar to areas previously described in squamous cell carcinomas of the cervix. Occasional squamous cell carcinomas of the cervix contain sheets of cells with abundant cytoplasmic glycogen and have been referred to as squamous cell carcinomas with glycogenated c y t ~ p l a s m . ~ ~ * ~ ~ An additional factor suggesting that these cases represent a variant of adenosquamous carcinoma was the finding of foci of spindle- shaped cells with abundant eosinophilic cytoplasm suggesting squamous differentiation in two cases. It should be emphasized, however, that none of the current cases contained keratin pearls, individual cell keratinization, or prominent intercellular bridges, and that in the ab- sence of clear-cut histologic squamous differentiation, we cannot unequivocally prove that these tumors represent adenosquamous carcinomas, rather than pure adenocarcinomas.

The cliff erential diagnosis of clear cell adenosquamous carcinoma includes clear cell carcinoma and glassy cell carcinoma, both of which can have areas composed of masses of cells with abundant cytoplasm. However, clear cell carcinoma and glassy cell carcinoma

have distinctive histologic or clinical features allowing them to be differentiated from clear cell adenosquamous carcinoma (Table 3). In addition to solid areas composed of cells with a vacuolated, glycogen-rich cytoplasm similar to those observed in clear cell adenosquamous carcinoma, clear cell carcinomas usually also have distinctive papillary and tubulocystic areas, More- over, cells with an eosinophilic granular cytoplasm and hobnail cells with scant cytoplasm and prominent nuclei that project into the lumens of the tubulocystic areas are characteristic of clear cell carcinoma. Both of these types of cells are not found in clear cell adenosquamous carcinoma. Moreover, unlike clear cell adenosquamous carcinoma, clear cell carcinomas are not associated with HPV-18, but instead, are associated with a history of intrauterine diethylstilbesterol exposure in two-thirds of cases.29

Although considered by some to be a rare variant of adenosquamous carcinoma, the most recent WHO classification of cervical cancers considers glassy cell carcinoma to be an entirely separate en tit^.'^,^' Clear cell adenosquamous carcinoma and glassy cell carcinomas are composed of solid masses of cells with prominent cellular borders. However, unlike the cells in clear cell adenosquamous carcinoma, the cells in glassy cell carcinomas are not vacuolated; instead, they have an abundant amphophilic or pale eosinophilic cytoplasm that has a characteristic ground glass appearance. Moreover, the cells in glassy cell carcinomas are characterized by large, oval nuclei with finely granular chromatin and with prominent nucleoli. In contrast, the nuclei of clear cell adenosquamous carcinoma are smaller, more irregular, have a more granular chromatin distribution, and usually lack prominent nucleoli.

1598 CANCER November 1, 1995, Volume 76, No. 9

Table 3. Characteristic Features of Clear Cell Adenosquamous Carcinoma

Clear cell adenosquamous Clear cell Glassy cell

Feature carcinoma carcinoma carcinoma

Solid areas Yes Yes Yes Focal glandular formations Yes Yes Yes

Hobnail cells No Yes No Vacuolated cytoplasm Yes Yes No Eosinophilic or granular cytoplasm No Yes Yes

Focal kera tinization No No Yes Large oval nuclei with prominent nucleoli No No Yes Distinct cell borders Yes No Yes Focal intracellular bridges No No Yes Exclusivelv associated with HPV 18 Yes No Unknown

Papillary or tubulocystic areas No Yes No

Focal mucin production Yes No YeS

HPV: human papillomavirus.

One of the most striking features of clear cell adenosquamous carcinoma of the cervix was the exclusive association of these tumors with HPV-18; all 11 of the cases included in this series were associated with HPV- 18. Although most recent large studies have detected HPV DNA in more than 80% of invasive squamous cell carcinomas, in most studies, HPV-16 is the predomi- nate HPV type associated with invasive squamous cell carcinoma of the cerv i~ . '~ ,~ ' - '~ Human papillomavirus- 18 is commonly found in association with other specific histologic types of invasive cervical cancer. One series of small cell undifferentiated carcinomas of the cervix detected HPV-18 in 14 (70%) of 20 cases3* Similarly, HPV-18 frequently is associated with invasive adenocarcinomas and adenosquamous carcinomas of the cervix (Table 4). Using in situ hybridization, Tase et al. de-

tected HPV-18 DNA in 8 (18%) of 44 adenosquamous carcinomas of the cervix. Using the polymerase chain reaction to identify HPV DNA, Yamakawa et al. detected HPV-18 DNA in 10 (48%) of 21 adenosquamous carcinomas of the cervix, and using Southern blot analysis, Walker et al. identified HPV-18 DNA in 3 (27%) of 11 adenosquamous carcinomas. However, in none of these series has the association between a specific type of HPV and a specific histologic type of invasive cervical cancer been as strong as the association between HPV- 18 and clear cell adenosquamous carcinoma.

Clear cell adenosquamous carcinoma appears to be a relatively uncommon neoplasm. The 11 cases reported here were identified from among 157 tumors originally classified as invasive cervical adenocarcinomas. However, despite the fact that it is uncommon, we

Table 4. Prevalence of Human Papillomavirus 16 and 18 in Adenocarcinomas and Adenosquamous Carcinomas of the Cervix

Adenosquamous Adenocarcinoma carcinoma

Author

% % Associated Associated w i t h HPV w i t h HPV

Detection No. of No. of method Cases 16 18 Cases 16 18

Tase et al. (1988)12 IHS 40 3 40 44 18 18

Walker et al. (1989)14 Southern blot 25 20 32 1 1 18 27 Wilczynski et al. (1988)13 Southern blot 11 18 45 -

Leminen et al. (1991)"' ISH 106 2 14 - Hording et al. (1991)9 PCR 50 18 52 - Das et al. (1993)' Southern blot 12 42 17 - Koulos et al. (1993)" PCR 32 31 28 -

Yamakawa et al. (1994)15 PCR 34 23 35 21 52 48 HPV: human papillomavirus; ISH: in-situ hybridization; PCR: polymerase chain reaction.

Clear Cell Adenosquamous Carcinoma/Fujiwara et al. 1599

believe that it is important to recognize this entity because of its poor prognosis. The overall tumor recurrence rate in the 11 women included in this series was 73%, and there was a 64% death rate. Even among women with Stage IB tumors, survival was poor. Three (60%) of the five women with Stage IB clear cell adenosquamous carcinoma died of their cervical disease. For comparison, the reported 5-year survival for women with Stage IB invasive cervical adenocarcinoma of all histologic subtypes is approximately 75% .39-41 Several years ago, it was suggested that cervical cancers associated with HPV-18 occurred in younger women, were more likely to be associated with lymph node involvement when classified as low stage, and had a worse prognosis than tumors associated with other types of ~ p v . 1 4 . 4 2 , 4 3 Although other series have not confirmed an association between HPV type and clinical outcome,33.34.44,45 the clear cell adenosquamous carcinomas described in this report occurred in relatively young women (mean age, 43 years), had aggressive clinical courses, and were associated with HPV-18. However, before concluding that these tumors have a particularly aggressive clinical course, additional studies are required comparing outcome of clear cell adenosquamous carcinoma with cervical cancers of a similar size and stage, but of a different histologic type.

References

1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

Kjaer SK, Brinton LA. Adenocarcinomas of the uterine cervix: the epidemiology of an increasing problem. Epidemiol Rev 1993; 35:486-98. Parazzini F, La Vecchia C. Epidemiology of adenocarcinoma of the cervix. Gynecol Oncol 1990;39:40-6. Shingleton HM, Gore H, Bradley D€i, Soong 5-J. Adenocarci- noma of the cervix. I.: clinical evaluation and pathologic features. A m 1 Obstef Gynecol 1981; 139:799-814. Tamimi HK, Figge DC. Adenocarcinoma of the uterine cervix. Gytiecol Oncol 1982; 13:335--44. Vesterinen E, Forss M, Nieminen U. Increase of cervical adenocarcinoma: a report of 520 cases of cervical carcinoma including 112 tumors with glandular elements. G y n ~ c o l Oncol 1989;33:49- 53. Chilvers C, Mant D, Pike MC. Cervical adenocarcinoma and oral contraceptives. BM] 1987; 295:1446-7. Devesa SS, Young JL, Brinton LA, Fraurneni JF. Recent trends in cervix uteri cancer. Cancer 1989;64:2184-90. Das BC, Gopakrishna V, Das DK, Sharma JK, Singh V, Luthra UK. Human papillomavirus DNA sequences in adenocarcinoma of the uterine cervix in Indian women. Cancer 1993;72:147-53. Hording U, Teglbjtirg CS, Visfeldt J, Bock JE. Human papillomavirus types 16 and 18 in adenocarcinoma of the uterine cervix. Gyiiecol Oncol 1992;46:313-6. Leminen A, Paavonen J, Vesterinen E, Wahlstrom T, Rantala I, Lehtinen M. Human papillomavirus types 16 and 18 in adenocarcinoma of the uterine cervix. A m 1 Clin Pafhol 1991;95:647- 52. Koulos J, Wright TC, Follen MM, Silva E, Atkinson EN, Richart

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

RM. Relationships between cKi-ras mutations, HPV types and prognostic indicators in invasive endocervical adenocarcinomas. Gynecol Oncol 1993;48:364-9. Tase T, Okagaki T, Clark BA, Manias DA, Ostrow RS, Twiggs LB, et al. Human papillomavirus types and localization in adenocarcinoma and adenosquamous carcinoma of the uterine cervix: a study by in situ hybridization. Cancer RES 1988;48:993-8. Wilczynski SP, Walker I, Liao SY, Bergen S, Berman M. Adeno- carcinoma of the cervix associated with human papillomavirus. Cancer 1988; 62:1331-6. Walker J, Bloss JD, Liao S-Y, Berman M, Bergen S, Wilczynslu SP. Human papillomavirus genotype as a prognostic indicator in carcinoma of the uterine cervix. Obstet Gytlecol 1989; 74:781-5. Yamakawa Y, Forslund 0, Teshima H, Hasumi K, Kitagawa T, Hansson BG. Human papillomavirus DNA in adenocarcinoma and adenosquamous carcinoma of the uterine cervix detected by polymerase chain reaction (PCR). Gytiecol Oncol 1994;53:190-5. Petterson F. Annual report on the results of treatment in gyne- cologic cancer. Stockholm: International Federation of Gynecol- ogy and Obstetrics, 1988. Happerfield LC, Bobrow LG, Bains R, Miller KD. Peroxidase la- belling immunocytochemistry: a comparison of eleven commer- cially-available avidin-biotin systems. Br ] Biomed Sci 1993; 50:

van den Brule AJC, Class ECJ, du Maine M, Melchers WJG, Helmerhorst 'T, Quint WGV, et al. Use of anticontamination primers in the polymerase chain reaction for the detection of human papillomavirus genotypes in cervical scrapes and biopsies. ] M e d Virol 1989;29:20-7. Maier RC, Norris HJ. Coexistence of cervical intraepithelial neoplasia with primary adenocarcinoma of the endocervix. Obstet Gynecol 1980;56:361-4. Dabbs DJ, Geisinger KR, Norris HT. Intermediate filaments in endometrial and endocervical carcinomas. Am J Surg Pafhol

Tamimi HK, Gown AM, Kim-Deobald J, Figge DC, Greer BE, Cain JM. The utility of immunocytochemistry in invasive adenocarcinoma of the cervix. Am ] Obstet Gynecol 1992; 166:1655-62. Kudo R, Sasano €1, Koizumi M, Orenstein JM, Silverberg SG. Immunohistochemical comparison of new monoclonal antibody IC5 and carcinoembryonic antigen in the differential diagnosis of adenocarcinoma of the uterine cervix. Int 1 Gynecol Pathol 1990; 9:325-36. Wahlstrom T, Lindgren J, Korhonen M, Seggala M. Distinction between endocervical and endometrial adenocarcinoma with immunoperoxidase staining of carcinoembryonic antigen in routine histological tissue specimens. Lancet 1979; 2:l 159-60. Kurman RJ, Noms HJ, Wilkinson E. Tumors of the cervix, vagina and vulva. 3rd series. Washington, DC: Armed Forces Institute of Pathology, 1992: 95. Scully RE, Bonfiglio TA, Kurman RJ, Silverberg SG, Wilkinson EJ. Histological typing of female genital tract tumors. 2nd ed. Berlin: Springer-Verlag, 1994: 46. Benda JA, Platz CE, Bushsbaum H, Lifshitz 5. Mucin production in defining mixed carcinoma of the uterine cervix: a clinicopath- ologic study. Int ] GynecoI Pathol 1985;4:314-27. Colgan TI, Auger M, McLaughlin JR. Histopathologic classification of cervical carcinomas and recognition of mucin-secreting squamous carcinomas. Int I G y n m l Pathol 1993; 1264-9. Wright TC, Ferenczy AF, Kurman RJ. Carcinoma and other tumors of the cervix. In: Kurman RJ, editor. Blaustein's pathology of the female genital tract. 4th ed. New York: Springer-Verlag, 1994: 279-326. Waggoner SE, Anderson SM, Eyck SV, Fuller J, Luce MC, Herbst

21-6.

1986; 10:568-76.

1600 CANCER November 2, 2995, Volume 76, No. 9

30.

31.

32.

33.

34.

35.

36.

AL. Human papillomavirus detection and p53 expression in clear-cell adenocarcinoma of the vagina and cervix. Obstet Gy- necol 1994; 84:404-8. Maier RC, Norris HJ. Glassy cell carcinoma of the cervix. Obstef Gynecol 1982;60:219-24. King LA, Tase T, Twiggs LB, Okagaki T, Savage JE, Adcock LL, et al. Prognostic significance of the presence of human papillomavirus DNA in patients with invasive carcinoma of the cervix. Cancer 1989;63:897-900. DeBritton RC, Hildesheim A, DeLao SL, Brinton LA, Sathya P, Reeves WC. Human papillomavirus and other influences on survival from cervical cancer in Panama. Obstet Gynecol 1993;81:

Higgins GD, Davy M, Roder D, Uzelin DM, Phillips GE, Burrell CJ. Increased age and mortality associated with cervical carcinomas negative for human papillomavirus RNA. Lancet 1991;338:

Jarrell MA, Heintz N, Howard P, Collins C, Badger G, Belinson J, et al. Squamous cell carcinoma of the cervix: HPV 16 and DNA ploidy as predictors of survival. Gynecol Oncol 1992; 46:361-6. Kenter GG, Cornelisse CJ, Jiwa NM, Aartsen EJ, Hermans J, Mooi W, et al. Human papillomavirus type 16 in tumor tissue of low-stage squamous carcinoma of the uterine cervix in relation to ploidy grade and prognosis. Cancer 1993;71:397-401. Lorincz AT, Reid R, Jenson AB, Greenberg MD, Lancaster W, Kurman RJ. Human papillomavirus infection of the cervix: relative risk associations of 15 common anogenital types. Obstet Gy- necoi 1992;79:328-37.

19-24.

910-3.

37.

38.

39.

40.

41.

42.

43.

44.

45.

Meijer CJLM, Van den Brule AJC, Snijders PJF, Kenemans P, Bleker OP, Walboomers JMM. Human papillomavirus and cervical cancer: the view of the gynaeco-pathologist. Cervix 1992; 10:101-8. Stoler MH, Mills SE, Gersell DJ, Walker AN. Small-cell neuro- endocrine carcinoma of the cervix: a human papillomavirus type 18-associated cancer. Am ] Surg Pathol 1991; 15:28-32. k m i n e n A, Paavonen J, Forss M, Wahlstrom T, Vesterinen E. Adenocarcinoma of the uterine cervix. Cancer 1990; 65:53-59. Saigo PE, Cain JM, Kim WS, Gaynor JJ, Johnson K, Lewis JL. Prognostic factors in adenocarcinoma of the uterine cervix. Can- cer 1986;57:1584-93. Eifel PJ, Morris M, Oswald J, Wharton IT, Delclos L. Adenocar- cinoma of the uterine cervix. Cancer 1990;65:2507-14. Barnes W, Delgado G, Kurman RJ, Petrilli ES, Smith DM, Ahmed S, et al. Possible prognostic significance of human papillomavirus type in cervical cancer. Gynecol Oncol 1988; 29:267-73. Kurman RJ, Schiffman MH, Lancaster WD, Reid 12, Jenson B, Temple GF, et al. Analysis of individual human papillomavirus types in cervical neoplasia: a possible role for type 18 in rapid progression. Am Obstet Gynecol 1988; 159:293-6. Kenter GG, Cornelisse CJ, Aartsen EJ, Mooy W, Hermans J, Heintz APM, et al. DNA ploidy level as a prognostic factor in low stage carcinoma of the uterine cervix. Gynecol Oncol 1990;39:181-5. Riou C, Favre M, Jeanne1 D, Bourhis J, Doussal VL, Orth G. As- sociation between poor prognosis in early-stage invasive cervical carcinomas and non-detection of HPV DNA. Lancet 1990;335: 1171-4.

clear cell adenosquamous carcinoma of the cervix · clear cell adenosquamous carcinoma of the...

Documents