To the Editor:
We read with interest the article “Eosinophils as a marker for invasion in cervical squamous neoplastic lesions” by Spiegel et al. (1). Their study on stromal eosinophils associated with invasion in cervical squamous neoplasms suggesting that eosinophils may be used as surrogate marker for invasion prompted us to study 58 cervical lesions similarly.
Thirty-two cases of squamous-cell carcinoma (30 of large cell nonkeratinizing type and 2 of keratinizing type) (Group I), 11 cases of cervical intraepithelial neoplasia (CIN) II and III (three with foci of microinvasion) (Group II), and 15 cases of chronic cervicitis with squamous metaplasia including two cases of CIN I (Group III) were analyzed for the presence and number of eosinophils as proposed by Speigel et al. (1). Leucocytes with brightly eosinophilic granules and bilobed nuclei were counted as eosinophils. Intravascular eosinophils were excluded from the counts. Scoring was performed at high dry magnification (×400) on Nikon Labophot 2 microscope (field area, 0.152 mm2).
Stromal eosinophils were present in 23 of 32 (71.8%), 3 of 11 (27.3%), and 4 of 15 (36.6%) cases in Groups I, II, and III, respectively; the 3 cases with eosinophils in Group II all had microinvasion. The differences between Groups I and II (p < 0.05) and between Groups I and III were statistically significant (p < 0.005). Stromal eosinophil scores were high in Group I, with up to 20 eosinophils/HPF. The scores for Group II and III were five eosinophils/HPF and two eosinophils/HPF respectively. Eosinophils ≥10 per 10 HPFs were seen in 19 of 23 (82.6%) cases of invasive carcinoma, 1 of 3 (33.3%) cases in Group II, and in none of the cases in Group III. The difference in stromal eosinophil density per 10 HPFs between Groups I and III was statistically significant (p < 0.005). The distribution of stromal eosinophils was focal and random. In cases of microinvasion, eosinophils were seen near the foci of invasion and occasional eosinophils infiltrated between tumor cells. Variable number of lymphocytes and plasma cells were present in all cases.
Tumor-associated eosinophilia has been reported in a wide variety of malignancies (2), including carcinoma of cervix (2,3). Spiegel et al. have, however, highlighted the significance of eosinophils in high-grade squamous intraepithelial lesions (HSILs). Their proposal that stromal eosinophil counts ≥5 per HPF and ≥10 per 10 HPFs in cervical biopsy specimens should alert pathologists to exclude invasive carcinoma is appropriate. In our study, we also found that ≥5 eosinophils per HPF and ≥10 eosinophils per 10 HPFs correlated well with invasion. All the cases of chronic cervicitis/CIN I had a stromal eosinophil density significantly lower than in the other groups. More importantly, within Group II (CIN II and III), only the 3 cases with microinvasion had ≥5 eosinophils per HPF. These observations are similar to those reported by Speigel et al. We also agree that the presence of stromal eosinophils in cases of CIN II and III should initiate a diligent search for invasion with deeper sections and/or additional biopsies.
The presence of eosinophils near areas of ulceration should be differentiated from stromal eosinophils. Increased number of stromal eosinophils have also been described in a variety of nonmalignant conditions such as parasitic infestations, topical applications, and prior surgical procedures (4). These associations should be excluded when evaluating the significance of stromal eosinophils in squamous neoplastic lesions of the cervix.
Sarla Agarwal, M.D., Professor
Neelam Wadhwa, M.D., Senior Research Associate
Geeta Gupta, M.D., Junior Resident
1. Spiegel GW, Asraf M, Brooks JJS. Eosinophils as a marker for invasion in cervical squamous neoplastic lesions. Int J Gynecol Pathol 2002; 21:117–24.
2. Lowe D, Jorizzo J, Hutt MSR. Tumor associated eosinophilia: a review. J Clin Pathol 1981; 34:1343–8.
3. Lowe DG. Carcinoma of cervix with massive eosinophilia. Br J Obstet Gynaecol 1988; 95:393–401.
4. Divack DM, Janovski NA. Eosinophilia encountered in female genital organs. Am J Obstet Gynecol 1962; 84:761–3.