Lymph node evaluation is an important step in the staging, prognosis and follow up of PCL, although there are differences in lymph nodal involvement between B-cell and T- cell lymphoma. Indeed, according to the TNM ISCL/EORTC staging system [
5,
6], in case of cutaneous B-cell lymphoma, microscopic evidence of regional lymph node involvement determines the transition from N0 to N1, N2 or N3 staging, depending on the peripheral region/s or central lymph node involvement respectively [
6]. In case of MF and SS, the evaluation of the lymph nodes is different; in fact, because of the negative impact on survival rates of “palpable adenopathy”, their clinical evidence alone determines the transition from N0 to N1 staging [
5]. This is a critical point in the management of MF/SS patients because the change in staging determines the need for systemic therapy; therefore, in clinical practice, histological evaluation can be requested. Criteria for lymph node removal in PCL staging are clinical and dimensional: the ISCL/EORTC revision [
5] defines as clinically abnormal peripheral nodes in MF/SS those measuring 1.5 cm or more in the longest transverse diameter, or any palpable peripheral node, regardless of size, that on physical examination is firm, irregular, clustered, or fixed [
5]. The 1.5-cm size is different from the 1-cm diameter node designated as abnormal in the ISCL/EORTC staging for non-MF/SS primary cutaneous lymphomas, since BRH or DL commonly occur in MF/SS, but arise less frequently in B-cell PCL [
6]. These enlarged or clinically abnormal lymph nodes should also be evaluated by ultrasonography (US), computed tomography (CT) and 18 F-fluorodeoxyglucose positron emission tomography (FDG-PET) [
5,
6]. In cases of multiple lymph adenopathies, the guidelines suggest the biopsy of the cervical or axillary nodes first, and then the inguinal ones [
5,
12], according to the different probabilities of involvement. As for the lymph node status, both the Dutch and NCI-VA classifications [
5] depended on their architecture and hence on histological evaluation. Therefore lymph node excision was the only considered procedure for this specific purpose. Nonetheless, simple histology too is not always effective; in the Fraser-Andrews’ study six of 19 patients with uninvolved lymph nodes or limited histological involvement (LN0-2) had a detectable T-cell clone at PCR investigation [
13]. Moreover excisional lymph node biopsies are not always easily performed and may be complicated by sepsis in immunodepressed, especially erythrodermic patients [
5]. Therefore the same guidelines suggest considering FNC as a possible diagnostic procedure for lymph node assessment possibly coupled with ancillary techniques [
5]. Notwithstanding, relatively few studies have investigated the possible contribution of FNC in lymph nodal investigation in PCL [
10,
11] probably because of the problem of lymph nodal FNC false negatives that may arise in other neoplasm. In fact FNC false negatives mainly concerns lymph nodes involved by neoplasm that determine partial and sub-capsular metastases, such as breast carcinoma and melanoma [
14,
15]. In these cases, the needle may not succeed in sampling the specific involved areas of the lymph node, producing false negatives. Nonetheless, lymph node partial involvements is less frequently observed in haematological neoplasm [
16] including cutaneous lymphoma. Moreover, according to the standardized cytological techniques of sampling, we moved the needle in different directions, during the FNC sampling, in order to reach different areas of the lymph node and to increase the probabilities of a representative sampling. Finally small cell clones may be not detected by FC but, in some CL, nuclear atypia are quite evident to be identified at the microscopic examination of the smears and by ICC, even though in small number. In the present study, although the basic approach of the technique was the same as that used in lymph nodal FNC from different contexts [
8], specific problems were encountered mainly due to the different sites of development and clinical presentation. In fact, it was either impossible, or hardly possible, to perform FNC under US control on axillary lymph nodes due to their “mobility” and to the anatomical peculiarity of the axillary hollow which does not provide enough room for the US probe and the lateral needle holder, or even for the needle cap to guide the needle. Therefore, in our experience, palpable axillary lymph nodes were better approached by “pulling down” the node with the index and middle fingers and blocking it on the thoracic wall. One of the fingers was then used as a support for the needle while carrying out the sampling with or without aspiration. The second problem was presented by erythrodermic patients who are generally more sensitive than others, and often suffer from cold and have fragile skin that easily bleeds. Therefore lymph nodal sampling was performed quite rapidly on these patients, with no more than one additional pass and with careful capitalization of the material. For the above mentioned reasons, in our Department, ROSE is generally performed on lymph node FNC and always on PCL patients. This methodology, in addition to assessing the adequacy of the smear, allows a correct management of the material obtained according to the cytological features it identifies. In our study, for instance, for small-medium sized cell populations, cell suspension in buffered solution for FC and PCR was generally preferred. Conversely, in case of large cell populations additional alcohol fixed smears were preferred for ICC phenotypization. Therefore we believe that the “petals” largely compensate for the “thorns” that ROSE offers in this contexts [
17]. In the study by Pai et al. [
11] the cell block technique was used for ICC; in our laboratory cell block is highly effective in different cytopathological contexts but, in this specific series, since we did not obtain sufficient cellularity in the first two cases, we preferred to utilize residual material for other ancillary techniques. As for the positive cases, the cytological features observed were quite similar to those described by Pai et al. [
11], and represent the only other extensive cytological description of lymph nodal MF/SS available, apart from the present one. However, we did not observe the predominant small-cell pattern as the one described by these authors. Ancillary techniques are basic tools in the cytological diagnosis of lymphoproliferative processes [
8,
9,
18‐
20]; according to Galindo [
10] and Pai [
11] they are fundamental in this specific context, whereas we believe that some points should be discussed. In our experience, according to Galindo et al. [
10] and Pai et al. [
11], FC by CD4/CD8 ratio and the quantitative evaluation of CD7 still seems to be the most effective procedures, provided that a sufficient amount of cells is collected. Even in the case of lymph node involvement by B-cell PCL, FC was effective in demonstrating the B-cell phenotype and the corresponding light chain restriction as in the corresponding primary NHL. Nonetheless, in the present series FC was not effective in 8 out of 21 cases; this performance was definitely less effective than that observed in other FNC/FC lymph nodal series [
8,
9] and other procedures had to be applied. This finding hampers the comparison of the different applied techniques and should be not be surprising too; in fact in case of large diagnostic cell populations as the case of small cell B or T-NHL the procedure is highly effective [
8,
9]; conversely when diagnostic cells are relatively scanty and intermingled among benign reactive cells FC is proportionally less effective [
8,
9]. This is the case of Hodgkin lymphoma or anaplastic large cells and some high-grade T-cell lymphoma in which diagnostic cells are too scanty to be gated or being too large, they stick to the tubes of the equipment or get broken or lost determining false negative results. In these cases, conversely even few cells detected on light microscope may be sufficient for a definitive diagnosis, therefore other ancillary techniques had to be used. For instance, in 3 cases of scanty cellularity, ICC on additional smears was more effective than FC, allowing the identification of large atypical T-cell cells that were too scanty to be analyzed by FC. As for molecular testing, TCR-PCR is the most sensitive technique to assess T-cell clonality but also carries a relatively high rate of false positives [
21]. In the study by Galindo et al. [
10] there was total agreement between cytological FC and TCR on cytological samples in terms of both sensitivity and specificity. Conversely, in the same study, tissue/TCR was highly sensitive but less specific, as it detected three reactive lymph nodes as clonal [
10]. Pai et al. [
11] detected T-cell clonal population by TCR-γ PCR in two cases and, as expected, negativity in two Hodgkin lymphoma. In our cases TCR-PCR was performed on cytological material only in 4 cases and was also in agreement with the cytological/FC data. In addition to Galindo’s experience [
10], a rate of false positive respectively of 3.6 and 5.4% for fresh and paraffin-embedded tissues was reported [
21]. In a study performed on peripheral blood of MF and SS patients a 34% positivity rate on patients with benign cutaneous infiltrate was detected [
22]. Therefore, as FNC is inevitably contaminated with peripheral blood, TCR-γ PCR should be carefully evaluated in this regard. Considering the dramatic evolution of molecular technologies, it is easy to foresee that highly sensitive and accurate procedures will shortly overcome specificity problems.