Clinicians and pathologists traditionally use patient data in addition to clinical examination to support their diagnoses.
We investigated whether a combination of histologic whole slides image (WSI) ...analysis based on convolutional neural networks (CNNs) and commonly available patient data (age, sex and anatomical site of the lesion) in a binary melanoma/nevus classification task could increase the performance compared with CNNs alone.
We used 431 WSIs from two different laboratories and analysed the performance of classifiers that used the image or patient data individually or three common fusion techniques. Furthermore, we tested a naive combination of patient data and an image classifier: for cases interpreted as ‘uncertain’ (CNN output score <0.7), the decision of the CNN was replaced by the decision of the patient data classifier.
The CNN on its own achieved the best performance (mean ± standard deviation of five individual runs) with AUROC of 92.30% ± 0.23% and balanced accuracy of 83.17% ± 0.38%. While the classification performance was not significantly improved in general by any of the tested fusions, naive strategy of replacing the image classifier with the patient data classifier on slides with low output scores improved balanced accuracy to 86.72% ± 0.36%.
In most cases, the CNN on its own was so accurate that patient data integration did not provide any benefit. However, incorporating patient data for lesions that were classified by the CNN with low ‘confidence’ improved balanced accuracy.
•Pathologists incorporate patient data in addition to clinical examination.•They put more emphasis on patient data if they are uncertain.•We investigated fusing histologic image/patient data within CNN-based classifiers.•State-of-the-art fusing approaches in general did not yield a performance benefit.•Mimicking humans by fusing patient data only if CNN was uncertain raised accuracy.
Cuticular poroma is a rare variant of poroma composed of exclusively or predominantly cuticular cells, namely of large cells with ample eosinophilic cytoplasm. We report 7 cases of this rare tumor ...identified among 426 neoplasms diagnosed as poroma or porocarcinoma. The patients were 4 males and 3 females, ranging in age from 18 to 88 years. All presented with a solitary asymptomatic nodule. The location included knee (2 cases), shoulder, thigh, shin, lower arm, and neck (each 1). All lesions were surgically removed. No evidence of disease was observed in 5 patients with available follow-up (range 12-124 months).Microscopically, all neoplasms were composed of variably sized, focally closed packed, or interconnecting nodules constituted mostly of cuticular cells. Small poroid cells were a focal feature in 5 tumors, whereas in the remaining 2 cases, poroid cells with conspicuous but still in minority. Five neoplasms were somewhat asymmetric, with irregular outlines. Ductal differentiation and intracytoplasmic vacuoles were seen in 6 tumors. Other features variably encountered were conspicuous intranuclear pseudoinclusions, cystic change, occasional multinucleated cells, increased mitoses, and stromal desmoplasia. Four of the 5 tumors analyzed with next-generation sequencing yielded YAP1::NUTM1 fusions. In addition, various mutations, mostly of unknown significance were identified in one neoplasm.
A basic requirement for artificial intelligence (AI)–based image analysis systems, which are to be integrated into clinical practice, is a high robustness. Minor changes in how those images are ...acquired, for example, during routine skin cancer screening, should not change the diagnosis of such assistance systems.
To quantify to what extent minor image perturbations affect the convolutional neural network (CNN)–mediated skin lesion classification and to evaluate three possible solutions for this problem (additional data augmentation, test-time augmentation, anti-aliasing).
We trained three commonly used CNN architectures to differentiate between dermoscopic melanoma and nevus images. Subsequently, their performance and susceptibility to minor changes (‘brittleness’) was tested on two distinct test sets with multiple images per lesion. For the first set, image changes, such as rotations or zooms, were generated artificially. The second set contained natural changes that stemmed from multiple photographs taken of the same lesions.
All architectures exhibited brittleness on the artificial and natural test set. The three reviewed methods were able to decrease brittleness to varying degrees while still maintaining performance. The observed improvement was greater for the artificial than for the natural test set, where enhancements were minor.
Minor image changes, relatively inconspicuous for humans, can have an effect on the robustness of CNNs differentiating skin lesions. By the methods tested here, this effect can be reduced, but not fully eliminated. Thus, further research to sustain the performance of AI classifiers is needed to facilitate the translation of such systems into the clinic.
•Convolutional neural networks (CNNs) show potential in skin cancer diagnoses.•Minor dermoscopic image changes suffice to influence the CNNs' diagnosis.•CNNs' susceptibility to such changes can be reduced but not eliminated.•Practitioners need to be aware and attempt to minimise this phenomenon.
Recent studies have shown that deep learning is capable of classifying dermatoscopic images at least as well as dermatologists. However, many studies in skin cancer classification utilize ...non-biopsy-verified training images. This imperfect ground truth introduces a systematic error, but the effects on classifier performance are currently unknown. Here, we systematically examine the effects of label noise by training and evaluating convolutional neural networks (CNN) with 804 images of melanoma and nevi labeled either by dermatologists or by biopsy. The CNNs are evaluated on a test set of 384 images by means of 4-fold cross validation comparing the outputs with either the corresponding dermatological or the biopsy-verified diagnosis. With identical ground truths of training and test labels, high accuracies with 75.03% (95% CI: 74.39-75.66%) for dermatological and 73.80% (95% CI: 73.10-74.51%) for biopsy-verified labels can be achieved. However, if the CNN is trained and tested with different ground truths, accuracy drops significantly to 64.53% (95% CI: 63.12-65.94%,
< 0.01) on a non-biopsy-verified and to 64.24% (95% CI: 62.66-65.83%,
< 0.01) on a biopsy-verified test set. In conclusion, deep learning methods for skin cancer classification are highly sensitive to label noise and future work should use biopsy-verified training images to mitigate this problem.
Background With the exception of erythema migrans, Borrelia infection of the skin manifests much more commonly with B cell–rich infiltrates. T cell–rich lesions have rarely been described. Objective ...We report a series of 6 patients with cutaneous borreliosis presenting with T cell–predominant skin infiltrates. Methods We studied the clinicopathologic and molecular features of 6 patients with T cell–rich skin infiltrates. Results Half of the patients had erythematous patchy, partly annular lesions, and the other patients had features of acrodermatitis chronica atrophicans. Histopathology revealed a dense, band-like or diffuse dermal infiltrate. Apart from small, well differentiated lymphocytes, there were medium-sized lymphocytes with slight nuclear atypia and focal epidermotropism. An interstitial histiocytic component was found in 4 cases, including histiocytic pseudorosettes. Fibrosis was present in all cases but varied in severity and distribution. In 5 patients, borrelia DNA was detected in lesional tissue using polymerase chain reaction studies. No monoclonal rearrangement of T-cell receptor gamma genes was found. Limitations This retrospective study was limited by the small number of patients. Conclusion In addition to unusual clinical presentation, cutaneous borreliosis can histopathologically manifest with a T cell–rich infiltrate mimicking cutaneous T-cell lymphoma. Awareness of this clinicopathologic constellation is important to prevent underrecognition of this rare and unusual presentation representing a Borrelia-associated T-cell pseudolymphoma.
Background
We present herein a series of 14 lesions showing overlapping features with the newly defined benign cutaneous mesenchymal neoplasm labeled as fibroblastic connective tissue nevus (FCTN).
...Methods and Results
Total of 8 patients were male and 5 were female, ranging in age from 1 to 56 years. Lesions appeared as isolated nodules or plaques on the trunk (7 cases), the limbs (4 cases) and the neck (2 cases). Histologically, all cases were composed of bundles of bland spindle cells of fibroblastic/myofibroblastic lineage irregularly branching within the reticular dermis and along fibrous septa in the subcutis. Adnexal structures and dermal adipocytes were entrapped by the fascicles, the epidermis was often papillomatous and elastic fibers were decreased and fragmented. Expression of CD34 and ASMA was found in 8 and 7 cases, respectively. Follow‐up was available for 7 patients (mean follow‐up, 5 years; range, 1‐10 years). None of the cases metastasized or recurred, even when incompletely excised.
Conclusion
The differential diagnosis of FCTN is broad and includes hypertrophic scar, dermatofibroma, dermatomyofibroma, pilar leiomyoma, plaque‐stage DFSP, CD34‐positive plaque‐like dermal fibroma, fibroblastic‐predominant plexiform fibrohistiocytic tumor, lipofibromatosis, superficial desmoid fibromatosis and fibrous hamartoma of infancy, of which it represents probably the monophasic variant.
Background A systematic analysis of the entire spectrum of various forms of differentiation and metaplastic epiphenomena in cutaneous apocrine mixed tumor (AMT) has never been performed. Objective ...The purpose of our study was to study a large number of cutaneous mixed tumors so as to fully characterize the entire spectrum of differentiations and metaplastic changes that may occur in the epithelial, myoepithelial, and stromal components of AMT. Methods This article reports a light-microscopic study of 244 cases of cutaneous AMT, complemented by a literature review. Results All types of differentiation along the lines of the folliculosebaceous-apocrine unit can be seen in AMT. The spectrum of metaplastic changes in the epithelial components includes squamous metaplasia, mucinous metaplasia, oxyphilic metaplasia, clear cell change, columnar metaplasia, hobnail metaplasia, and cytoplasmic vacuolization. The following changes in the myoepithelial component were documented: clear cell change, hyaline cells, plasmacytoid cells, spindling, and collagenous spherulosis. Stromal alterations included chondroid metaplasia, osseous metaplasia, and adipose metaplasia. Limitations This study utilizes tissue specimens that mainly came as consultations; therefore some inherent selection bias exists. Conclusions AMT displays a wide range of differentiation and metaplastic changes in its epithelial, myoepithelial, and stromal components. These phenomena are not mutually exclusive. When unduly prominent, they may present diagnostic pitfalls. Our findings corroborate those of previous publications, stressing the remarkable diversity of differentiation and metaplasias that may be found in cutaneous AMT. We propose that the most appropriate name for these lesions is “mixed tumor of the folliculosebaceous-apocrine complex.”
We report 50 cases of peculiar histiocytic proliferations occurring in diverse body sites that currently bear various names, including nodular mesothelial/histiocytic hyperplasia, nodular histiocytic ...aggregates, mesothelial/monocytic incidental cardiac excrescences, reactive eosinophilic pleuritis, histioeosinophilic granuloma of the thymus, and intralymphatic histiocytosis. They can sometimes cause considerable differential diagnostic difficulties by resembling a metastatic carcinoma or Langerhans cell histiocytosis. Several previous publications have established a link between some of these conditions, suggesting that these are merely variations within a histopathologic spectrum, affecting different organs and bearing different names based on a particular location. However, no publication has ever comprehensively addressed all of these lesions together in one study in an attempt to explain and discuss their striking analogy. Having studied a large series of cases we provide evidence that all these lesions share the same morphologic, immunohistochemical, and pathogenetic properties, thus they all represent the same pathologic process and should be referred to as such. Taking into account their typical nuclear features we propose a collective term “histiocytosis with raisinoid nuclei” for this spectrum of conditions.
In our routine and consultative pathology practices, we have noticed that a relatively high proportion of spindle cell predominant trichodiscomas demonstrate a remarkable stromal admixture of adipose ...tissue, which along with spindle cells, prominent collagen bundles and myxoid change closely resembles spindle cell lipoma (SCL). To clarify their possible relationship to SCL, 25 cases of trichodiscoma and fibrofolliculoma with stromal "lipomatous metaplasia" were collected and examined using immunohistochemical stains CD34 and retinoblastoma-1 (RB1) protein and fluorescence in situ hybridization (RB1 deletion). The patients ranged in age from 35 to 81 years (median 64 years). The male to female ratio was almost equal (14:11). All tumors with a known location were situated on the face with a special predilection for the nose. All cases were sporadic, with all patients having a single lesion and showing no clinical features of Birt-Hogg-Dubé syndrome. No case with available follow-up presented with a recurrence or an otherwise aggressive clinical course. Spindle cell stroma was immunohistochemically positive for CD34 in 16 of 20 cases, and 18 of 19 cases showed loss of RB1 staining in lesional spindle cells. Fluorescence in situ hybridization analysis detected RB1 gene heterozygous deletion in 6 of 20 cases. We conclude that despite the SCL-like appearance of the investigated cases, the majority of them supposedly represent genuine spindle cell predominant trichodiscomas with adipose tissue admixture. However, there was a subset of histopathologically indistinguishable cases with proved RB1 deletion, which likely represent SCL with trichodiscoma/fibrofolliculoma-like epithelial/adnexal induction rather than spindle cell predominant variant of trichodiscoma.
Overdiagnosis of melanoma – causes, consequences and solutions Kutzner, Heinz; Jutzi, Tanja B.; Krahl, Dieter ...
Journal der Deutschen Dermatologischen Gesellschaft,
November 2020, 2020-Nov, 2020-11-00, 20201101, Letnik:
18, Številka:
11
Journal Article
Recenzirano
Summary
Malignant melanoma is the skin tumor that causes most deaths in Germany. At an early stage, melanoma is well treatable, so early detection is essential. However, the skin cancer screening ...program in Germany has been criticized because although melanomas have been diagnosed more frequently since introduction of the program, the mortality from malignant melanoma has not decreased. This indicates that the observed increase in melanoma diagnoses be due to overdiagnosis, i.e. to the detection of lesions that would never have created serious health problems for the patients. One of the reasons is the challenging distinction between some benign and malignant lesions. In addition, there may be lesions that are biologically equivocal, and other lesions that are classified as malignant according to current criteria, but that grow so slowly that they would never have posed a threat to patient’s life. So far, these “indolent” melanomas cannot be identified reliably due to a lack of biomarkers. Moreover, the likelihood that an in‐situ melanoma will progress to an invasive tumor still cannot be determined with any certainty. When benign lesions are diagnosed as melanoma, the consequences are unnecessary psychological and physical stress for the affected patients and incurred therapy costs. Vice versa, underdiagnoses in the sense of overlooked melanomas can adversely affect patients’ prognoses and may necessitate more intense therapies. Novel diagnostic options could reduce the number of over‐ and underdiagnoses and contribute to more objective diagnoses in borderline cases. One strategy that has yielded promising results in pilot studies is the use of artificial intelligence‐based diagnostic tools. However, these applications still await translation into clinical and pathological routine.