Paolo Greco

Adenocarcinoma of the Lower Third of the Rectum Surgically Treated With a <10-MM Distal Clearance: Preliminary Results in 35 N0 Patients

Background: Recent reports suggest that a distal clearance (DC) of 10 mm at the lower surgical margin may be considered adequate in the surgical treatment of rectal cancer, but there are no data on the possible adequacy of a <10-mm DC in N0 patients in whom a good prognosis can otherwise be expected, that is, those with negative surgical margins and negative lymph nodes.Methods: Between November 1991 and December 1998, 154 consecutive patients with adenocarcinoma of the lower third of the rectum had a total rectal resection with total mesorectal excision and coloendoanal anastomosis. Among 76 N0 patients, there were 35 with <10-mm DC and 41 with ≥10-mm DC. Each group was divided into two subgroups depending on whether the surgical margins were involved or not, and the rate of local recurrence in the various categories was compared. All B2 Astler-Coller stage patients in the series received postsurgical chemoradiotherapy.Results: The local recurrence rate in the 35 patients with DC <10 mm was 11.4% and that of the 41 patients with DC ≥10 mm was 7.3%. When only patients with negative surgical margins were considered, the local recurrence rate was 3.4% for those with <10-mm DC and 5.1% for those with ≥10-mm DC.Conclusions: Our results suggest that a radical surgery with <10-mm DC followed by chemoradiotherapy may be adequate in N0 patients, provided that a careful pathologic examination of the surgical specimen excludes the presence of lymph node metastases and that the distal rectal and mesorectal resection margins fall in healthy tissue.

Solitary fibrous tumour of the kidney with sarcomatous overgrowth

Solitary fibrous tumour (SFT) rarely occurs in the kidney, with only one case exhibiting malignant behaviour. We report the case of a typical SFT of the kidney with sarcomatous overgrowth in a 34‐year‐old woman. This malignant component, grossly apparent as a nodular area arising in the context of the main tumour mass, consisted of CD34+ mitotically active atypical plump spindle‐ to epithelioid‐shaped cells, including pleomorphic multinucleated giant cells. A novel immunohistochemical finding was diffuse and strong S‐100 protein expression by sarcomatous cells. This should be kept in mind by pathologists to avoid confusion with other S‐100 protein‐positive malignant neoplasms.

Deciduoid-like myofibroblastoma of the breast: a potential pitfall of malignancy.

Sir: Mammary stroma plays a crucial role in regulating breast epithelial growth, morphogenesis, hormone responsiveness and neoplastic progression. It is also well known that mammary stromal cells are endowed with plasticity to undergo changes in phenotype, as a result of a metaplastic or reactive process. Typical examples are the myofibroblastic or leiomyomatous metaplasia that can occur in fibroadenomas, phylloides tumours, myofibroblastomas, or in basically similar benign lesions, variously labelled with the terms hamartomas, mesenchymoma, stromo-epithelial or fibro-epithelial proliferations. In some cases of hamartoma and myofibroblastoma additional heterologous mesenchymal components, such as mature fat, cartilage, bone or even skeletal muscle may coexist. Apart from metaplastic processes, mammary stromal cells may undergo reactive morphological changes in several local or systemic pathological conditions, such as post-fine needle aspiration reactive spindle cell nodules, diabetes mellitus, lupus erythematosus and hypothyroidism. In the last three sytemic diseases, stromal cells may acquire an epithelioid to deciduoidlike morphology with worrisome features that can be confused with infiltrating carcinoma. Recently, we have encountered a case of mammary myofibroblastoma (MFB) composed exclusively of deciduoid-like cells similar to those previously reported by us in non-neoplastic conditions. We think that the tumour presented here is a previously unrecognized morphological variant of MFB, for which the descriptive term ‘deciduoid-like MFB’ is proposed. A 47-yearold woman presented with a painless, freely mobile solitary, 20-mm lump in her left breast. Clinical history was unremarkable. Preoperative ultrasonography and mammography confirmed a circumscribed, solid tumour mass in the breast parenchyma, consistent with fibroadenoma. Complete surgical excision of the mass, including a rim of adjacent grossly normal parenchyma, was performed. No recurrence has been experienced 6 months after surgery. Grossly, a wellcircumscribed, unencapsulated tumour mass, 20 mm in diameter, was seen. On cut section, the nodular mass was whitish in colour and firm in consistency. Histological examination revealed a well-circumscribed neoplasm composed exclusively of densely packed large cells with a solid or trabecular growth pattern, focally arranged in nests (Figure 1A). These cells were round to ovoid to polygonal in shape, with abundant eosinophilic glassy cytoplasm and sharp cellular borders (Figure 1B). Nuclei were large and round, with vescicular chromatin containing a single or multiple prominent nucleoli (Figure 1B). The cells had an overall appearance reminiscent of decidua (Figure 1B) and were named ‘deciduoid-like cells’, accordingly. Numerous cells were binucleated. Focally, some cells, showing eccentric nuclei and spherically eosinophilic intracytoplasmic inclusions were reminiscent of rhabdoid cells (Figure 1B). The mitotic index was low [one mitosis · 10 high-power fields (1HPF = 0.159mm)]. Atypical mitoses, necrosis or haemorrhage were absent. Thick eosinophilic collagen fibres, sometimes with an amianthoid-like appearance, were frequently observed among cells or around cellular nests (Figure 1A). Mammary ducts or lobules were not trapped within the tumour. Immunohistochemistry revealed diffuse, intense immunoreactivity to vimentin, a-smooth muscle actin, desmin (Figure 1C), CD34, Bcl-2, CD99, oestrogen (100%), progesterone (90%) and androgen (90%) receptors (Figure 1D). H-caldesmon was only focally detected (<1% of neoplastic cells). No immunoreactivity was obtained with pancytokeratins, epithelial membrane antigen, CD10, calponin, S100 protein, HMB-45, CD68, p63, myogenin or MyoD-1. This case appears unique in that, being composed exclusively of deciduoid-like cells, it was difficult to recognize as MFB. The differential diagnosis included both malignant and benign tumours, such as apocrine carcinoma, melanoma, pleomorphic rhabdomyosarcoma, epithelioid leiomyoma, epithelioid angiomyolipoma and epithelioid schwannoma. However, due to the site of origin of the tumour, the most important differential diagnosis revolved around apocrine carcinoma. Our case looked like apocrine carcinoma in that neoplastic cells were large, with abundant eosinophilic cytoplam and large vesicular nuclei containing prominent nucleoli. Although the absence of a high mitotic index, atypical mitoses, necrosis, and infiltrative growth pattern argued against malignancy, immunohistochemistry was extremely helpful in ruling out apocrine carcinoma, revealing myofibroblastic differentiation of neoplastic cells, along with no expression of epithelial markers. Melanoma was easily excluded due 652 Correspondence

Warthin tumor-like papillary thyroid carcinoma with a minor dedifferentiated component: report of a case with clinicopathologic considerations.

Warthin tumor-like papillary thyroid carcinoma is an uncommon variant of papillary thyroid carcinoma. We report a rare case of Warthin tumor-like variant of papillary thyroid carcinoma with a dedifferentiated component consisting of a solid tumor area composed of neoplastic cells with a spindle to tall cell morphology associated with marked nuclear pleomorphism, atypical mitoses, and foci of necrosis. Although our patient presented with a locally aggressive disease (T3 N1b Mo), she is disease-free without radioiodine therapy after a 23-month follow-up period. We emphasize that Warthin tumor-like papillary thyroid carcinoma, like other morphological variants of papillary carcinoma, may occasionally undergo dedifferentiation. As this component may be only focally detectable, we suggest an extensive sampling of all large-sized (>3 cm) papillary thyroid carcinoma. Recognition of any dedifferentiated component in a Warthin tumor-like papillary thyroid carcinoma should be reported, including its percentage, because it may reflect a more aggressive clinical course.

Deciduoid-like stromal cells in a diabetic patient with bilateral gynecomastia: a potential diagnostic pitfall

Sir, Epithelioid stromal cells (ESCs) were originally described as a typical feature of diabetic mastopathy, along with lymphocytic mastitis [5]. As these cells have also been reported in the breast stroma of patients with hypothyroidism and systemic lupus erythematous, it has been postulated that they may represent a reliable morphological marker of autoimmune disorders [1]. Diabetic mastopathy may be clinically and histologically confused with carcinoma [2], especially in those cases where there is a presence of large-sized ESCs [1]. We report the first case of gynecomastia with extensive deciduoid-like stromal changes in a patient affected by long-standing insulin-dependent diabetes mellitus (IDDM). A 44-year-old man presented with a retro-areolar nodule in the left breast. Ultrasound analysis revealed an irregular outlined, non-homogeneous hypoechoic mass. The lesion was surgically excised. After a 1-year clinical follow-up, a similar lesion was noted and excised from the contralateral breast. The patient referred to had a 31-year history of IDDM, and he had been undergoing hemodialysis for diabetic nephropathy for 2 years. Grossly, both surgical samples consisted of unencapsulated fibrous nodular masses, measuring 3 cm and 4 cm in greatest diameter, respectively, with a firm, grayish-white cut surface. Histologically, both lesions exhibited the typical features of florid gynecomastia: increased number of ducts with epithelial proliferation and periductal myxoid stroma containing slight lymphocytic infiltration (Fig. 1A). An unexpected feature was the presence in the interlobular stroma of numerous, large cells with deciduoid-like features, arranged in solid sheets and embedded in fibrous tissue (Fig. 1A, B). These cells were round, ovoid, or polygonal in shape, with an abundant pale-to-eosinophilic cytoplasm, well-defined cellular borders and large, round vesicular nuclei containing one or more prominent nucleoli (Fig. 1B). A low number of cells were binucleate and showed a granular eosinophilic cytoplasm. A mild nuclear pleomorphism was focally seen. Mitotic figures were seen infrequently (2–3 50 high-power field). No atypical mitosis or necrosis was identified. A moderate stromal perivascular lymphocytic infiltrate was observed. Immunohistochemical analyses, revealing reactivity for vimentin, a-smooth-muscle actin (20–30% of cells) (Fig. 1C) and no staining with pancytokeratins, epithelial membrane antigen, desmin, h-caldesmon, calponin, CD34, S100 protein, HMB45, CD68 and p63 were consistent with a fibro/myofibroblastic nature of deciduoid-like stromal cells. The association of diabetes mellitus with gynecomastia has only rarely been reported [3]. As plump stromal cells are lacking in common gynecomastia [3], we think that the deciduoid-like stromal cells of the present case represent a morphological variant of ESCs, typically occurring in diabetic mastopathy [1, 5]. It is likely that the deciduoid-like morphology is the result of reactive stromal changes to a hormonally related proliferative disease (gynecomastia) and to hemodialytic treatment, which may play a role through tissue micro-environmental modifications. In a large series of lymphocytic mastitis due to different basic diseases, the two most florid examples of epithelioid stromal changes were seen in the only diabetic male patient and in a woman undergoing hemodialysis by systemic lupus erythematosus [1]. Interestingly, our patient was a diabetic man undergoing hemodialysis. Awareness of deciduoid-like stromal changes in patients with the association diabetes mellitus/gynecomastia is crucial for pathologist to avoid confusion with benign and malignant tumors. Differential diagnosis mainly reG. Magro ()) Dipartimento G.F. Ingrassia, Anatomia Patologica, Universit di Catania, Via S. Sofia 87, 95123 Catania, Italia e-mail: g.magro@unict.it Tel.: +39-95-3782024 Fax: +39-95-3782023

Comments on the reporting of lymph nodes and tumour budding in the checklist of colorectal carcinoma

Dear Sir: We have read with interest the paper of Jass et al. [3] about the recommendations for the reporting of surgically resected specimens of colorectal carcinoma. We would like to comment on two controversial issues dealing with the number and the size of lymph nodes to be recovered and with the reporting of tumour budding. Although we admit that there is no unacceptable low number of lymph nodes for an individual dissection, we do not agree with the authors when they state that 12–15 (lymph nodes) is the mean number giving an indication of good practice in a series of colorectal cancer dissections [3]. There is no general consensus about this number if other authors, recently reporting colorectal cancer guidelines, propose a mean number of 15–18 lymph nodes [6]. Recently, we have published in Virchows Archiv a paper showing that in pT3 rectal cancer, 12–15 lymph nodes are insufficient for a correct staging [2]. In that prospective study, we first showed that if the mean number of lymph nodes recovered increased from 17.8 to 26.8, after a second sampling of the same surgically resected specimens, metastases were detected in 18.7% of patients that had been originally understaged [2]. Our results are in line with those previously obtained by other authors showing that most lymph-node metastases (about 80–90% of cases) of colorectal cancer are detected if about 26–30 lymph nodes are recovered [2]. Jass et al. do not provide any comment about the possibility that some patients with colorectal cancer have only a single metastatic lymph node. In a large series of colorectal cancers reported by Goldstein [1], about 30% of pT3N+ patients had a single metastatic lymph node, and the percentage of specimens with one lymph-node metastasis increased from 41.62 to 80.36% when the number of lymph nodes increased from 11–15 to more than 21, respectively. Notably, in our study, we obtained similar results, as all restaged pT3N+ patients had only a single metastatic lymph node, which was found in specimens in which a minimum number of 19 lymph nodes had been originally detected [2]. Accordingly, we propose that at least 20 lymph nodes should be recovered as an acceptable mean number in a series of colorectal cancer dissections. Apart from the number, in our opinion, the size of the lymph nodes to be recovered is crucial for a correct staging of colorectal cancer. It is known that a significative number of metastatic lymph nodes are small in size and can be easily missed during manual dissection. This is confirmed by our previously mentioned study showing that all restaged pT3N+ patients had a single metastatic lymph node measuring less than 5 mm in greatest diameter [2]. Another important issue refers to tumour budding. The authors state that tumour budding should be merely reported as present or absent [3]. Although we generally agree in reporting tumour budding, however, we point out that there is no general consensus about the definition and quantification. While Jass et al. regard tumour budding as single isolated cancer cells or a cluster composed of four cells at the invasive front [3], others include in the definition a cluster composed of five cells or even generic small clusters (number of cells not established) of undifferentiated cancer cells, also known as solitary trabecular forms [5]. There is no general consensus even about the Virchows Arch (2007) 450:359–360 DOI 10.1007/s00428-007-0366-y

Epithelial misplacement in the muscularis propria after biopsy of a colonic adenoma

Sir, Hyper-plastic and hamartomatous colonic polyps may show misplaced epithelium (pseudo-invasion) within the submucosa or more rarely even extending into the muscular wall and/or serosa, potentially simulating an invasion by a well-differentiated adenocarcinoma [2, 4]. The pathogenesis of this condition is referred to as the passage of the mucosal glands through the anatomical defects in the muscularis mucosae due to tissue damage from torsion, or twisting, of the polyp [2, 4]. It is likely that even minor trauma, resulting from vigorous peristalsis or bulky intraluminal material, may be responsible of epithelial misplacement, especially in sessile polyp [4]. Pseudo-invasion can also commonly occur in peduncolated adenomatous polyps, especially in the form of herniated mucosal glands through natural spaces in the muscularis mucosae, into the sub-mucosa of the polyp stalk [3]. Although it is known that transient epithelial misplacement in the sub-mucosa can occur in colonic adenomas within a short period after biopsy [1], to the best of our knowledge, an extension into the muscularis propria has not been reported to date. We first report an epithelial misplacement (pseudoinvasion) in the muscularis propria after an endoscopic biopsy of a sessile colonic adenoma. An 86-year-old man underwent colonoscopy for rectal bleeding. A large broad-based polypoid mass was found in the sigmoid colon, and bioptic material was obtained for histological examination. Histologically, the removed fragments were consistent with an adenoma with low-grade dysplasia. Because of the large dimension of the tumour, surgical resection (sigmoid colon and upper rectum) was performed after 12 days from biopsy. Grossly, a broad-based polypoid mass, measuring 4.5 cm in diameter, was observed in the sigmoid colon. The tumour was entirely sampled and processed for microscopic examination. Histologically, the tumour was a tubulovillous adenoma with low-grade dysplasia. At the base of the tumour, mucosal ulceration covered by a mixture of mucus lakes and suppurative exudate with isolated free-lying glands could be noted (Fig. 1a). It is surprising to note that isolated atypical glands were found in the underlying muscularis propria (Fig. 1b). Cytologically, the glandular epithelium was composed of atypical mucin-depleted cells with enlarged, vesicular, oval to round nuclei containing small nucleoli. These glands were completely enveloped by granulation tissue containing inflammatory cells, newly formed capillaries and a minor amount of fibroblasts (Fig. 1b). Despite additional histological sections prepared from deeper levels of the tissue blocks, foci of malignant transformation were not identified. It was possible to identify a continuous rim of granulation tissue in only one section, extending from the mucosal ulceration at the base of the tumour into the sub-mucosa and muscularis propria. Isolated atypical glands, cytologically similar to those formerly observed, were entrapped at different levels (from mucosa to muscularis propria) within the abovementioned granulation Virchows Arch (2007) 450:603–605 DOI 10.1007/s00428-007-0399-2

Tumour deposits classification in colorectal cancer: is TNM5 better than TNM7?†

We read with great interest the Invited Commentary ‘Evidence-based medicine: the time has come to set standards for staging’ [1] and the Invited Response ‘Evidence-based medicine: the time has come to set standards for staging. Is a radical overhaul really needed?’ [2]. We would like to comment on some controversial issues concerning the classification of tumour deposits (TDs) in colorectal cancer and the utilization of the fifth rather than the seventh edition of the tumour, nodes, metastasis (TNM) classification. Regarding the dimensional criterion used in TNM5 about TD classification, we do not agree with the statement by Quirke et al that ‘advantages were reproducibility and correlating with imaging’ [1]. Size, in fact, is not informative on the histogenesis of TDs. A TD ≤3 mm may be a totally metastatic lymph node; on the contrary, a TD >3 mm may represent vascular invasion with extravascular tumour growth. Therefore we believe that size may not represent a valid criterion to label TD as lymph node metastasis. Goldstein et alshowed that palpable pericolonic TDs, which according to TNM5 should be classified as metastatic lymph nodes, represented perineural, perivascular or intravascular tumour growth and there was no evidence of lymph node metastasis upon step sectioning [3]. In addition, a recent study showed that even small TDs (<2 mm), which should be included in the T-stage according to TNM5, contributed to poor prognosis [4]. Therefore it is crucial not to apply a dimensional or morphological criterion (size and/or shape) to prove which TDs are true metastatic lymph nodes, but to establish whether TDs have a behaviour similar to or more aggressive than that of lymph node metastases. Regarding the statement by Quirke et al [1] about differential diagnosis between TDs and tumour budding (TB), although there is no general consensus about TB definition and quantification, we doubt that pathologists could misdiagnose them as TDs [5]. In a recent study, TB was defined as a single or a group of <5 detached tumour cells at the invasive tumour front [6]. On the contrary, TDs are defined in TNM7 as “discrete foci of tumour found in the pericolic or perirectal fat or in adjacent mesentery (mesocolic fat) away from the leading edge of the tumour, showing no evidence of residual lymph node tissue but within the lymph drainage area of primary carcinoma . . .” [7,8]. In our opinion, the difference between the two types of lesions is evident. We agree with Quirke et al that TD definition after radio-chemotherapy was not addressed in TMN7 [1]. However, this issue was not resolved even in TNM5. Regarding TNM7, it is unclear why TDs should be classified as N1c in otherwise node-negative cases of T1–T2 colorectal cancer, similarly with TNM staging of skin melanoma [2,7]. There is no clinical or biological evidence indicating why satellite nodules found in melanoma and TDs found in colorectal cancer should be considered similar lesions. It is also unclear why the interpretation of TDs in T3–T4 colorectal cancer is left to the individual pathologist, reverting to TNM4 staging, although recent data show that all patients with TDs have a prognosis similar to or even worse than that of patients with nodal metastases [3,4,9]. Ueno et al, using the Akaike Information Criteria, showed that the best prognostic model predictive of survival outcome is obtained by including extravascular TDs in the N-stage and intravascular TDs in the T-staging [9]. If these data were to be confirmed, then all TDs should be included in the N-stage, regardless of T-stage, size and shape, except for TDs defined as ‘vascular type’ or intravascular TDs, which, in our opinion, would be better classified as extramural vascular (lymphatic or venous) invasion [10]. Despite this critical analysis of TNM7, we believe that the use of this edition is mandatory because there is no evidence that TNM5 is better than TNM7. In addition, N-stage should not be based on the personal decision of a pathologist to adhere to TNM5 or TNM7 because the use of two different classifications creates poor reproducibility and stage migration. In daily practice, we believe that TDs should be included in the pathology report, specifying their total number, size (the largest diameter if there are multiple lesions), and association, or not, with large vessels and/or nerves, in order to create more homogeneous groups of patients for enrolment in clinical trials [5,11]. This is crucial to better define the actual prevalence and clinical implication of TDs.

Cavernous hemangioma of the colon and rectum: a case report.

A case of cavernous hemangioma of the colon and rectum in a 21-year-old male is described. Histology showed an extensive vascular hamartoma. Intestinal hemangiomas may be difficult to diagnose mainly because of their rarity, but they are an important cause of bleeding.