Scoring of tumor-infiltrating lymphocytes: From visual estimation to machine learning.

The extent of tumor-infiltrating lymphocytes (TILs), along with immunomodulatory ligands, tumor-mutational burden and other biomarkers, has been demonstrated to be a marker of response to immune-checkpoint therapy in several cancers. Pathologists have therefore started to devise standardized visual approaches to quantify TILs for therapy prediction. However, despite successful standardization efforts visual TIL estimation is slow, with limited precision and lacks the ability to evaluate more complex properties such as TIL distribution patterns. Therefore, computational image analysis approaches are needed to provide standardized and efficient TIL quantification. Here, we discuss different automated TIL scoring approaches ranging from classical image segmentation, where cell boundaries are identified and the resulting objects classified according to shape properties, to machine learning-based approaches that directly classify cells without segmentation but rely on large amounts of training data. In contrast to conventional machine learning (ML) approaches that are often criticized for their "black-box" characteristics, we also discuss explainable machine learning. Such approaches render ML results interpretable and explain the computational decision-making process through high-resolution heatmaps that highlight TILs and cancer cells and therefore allow for quantification and plausibility checks in biomedical research and diagnostics.

A gene signature to predict high tumor-infiltrating lymphocytes after neoadjuvant chemotherapy and outcome in patients with triple-negative breast cancer.

Background: In patients with triple-negative breast cancer (TNBC), the extent of tumor-infiltrating lymphocytes (TILs) in the residual disease after neoadjuvant chemotherapy (NACT) is associated with better prognosis. Our objective was to develop a gene signature from pretreatment samples to predict the extent of TILs after NACT and then to test its prognostic value on survival. Patients and methods: Using 99 pretreatment samples, we generated a four-gene signature associated with high post-NACT TILs. Prognostic value of the signature on distant relapse-free survival (DRFS) was first assessed on the training set (n = 99) and then on an independent validation set (n = 115). Results: A four-gene signature combining the expression levels of HLF, CXCL13, SULT1E1, and GBP1 was developed in baseline samples to predict the extent of lymphocytic infiltration after NACT. In a multivariate analysis performed on the training set, this signature was associated with DRFS [hazard ratio (HR): 0.28, for a one-unit increase in the value of the four-gene signature, 95% confidence interval (CI): 0.13-0.63)]. In a multivariate analysis performed on an independent validation set, the four-gene signature was significantly associated with DRFS (HR: 0.17, 95% CI: 0.06-0.43). The four-gene signature added significant prognostic information when compared with the clinicopathologic pretreatment model (likelihood ratio test in the training set P = 0.004 and in the validation set P = 0.002). Conclusions: A four-gene signature predicts high levels of TILs after anthracycline-containing NACT and outcome in patients with TNBC and adds prognostic information to a clinicopathological model at diagnosis.

Genome-wide gene expression profiling to predict resistance to anthracyclines in breast cancer patients.

Validated biomarkers predictive of response/resistance to anthracyclines in breast cancer are currently lacking. The neoadjuvant Trial of Principle (TOP) study, in which patients with estrogen receptor (ER)-negative tumors were treated with anthracycline (epirubicin) monotherapy, was specifically designed to evaluate the predictive value of topoisomerase II-alpha (TOP2A) and develop a gene expression signature to identify those patients who do not benefit from anthracyclines. Here we describe in details the contents and quality controls for the gene expression and clinical data associated with the study published by Desmedt and colleagues in the Journal of Clinical Oncology in 2011 (Desmedt et al., 2011). We also provide R code to easily access the data and perform the quality controls and basic analyses relevant to this dataset.

Breast cancer response to paclitaxel in vivo.

Clinical trials using neoadjuvant (primary, preoperative) chemotherapy demonstrate that breast cancer reduction relates to survival. To date, no pre-treatment pathologic, phenotypic, or genotypic tumor characteristics predict a patient's likely benefit from paclitaxel. This has led to pilot clinical studies that have attempted to identify whether early cellular responses in vivo can be used to predict the effectiveness of chemotherapy. A potential benefit of such predictive studies will be the ability to tailor specific therapeutic approaches to individual patients. Important issues surrounding this field include how to accurately measure and/or categorize the extent of tumor reduction, and how and when to assess breast cancer cellular responses in vivo. Preliminary data indicate that initial apoptotic responses are critical to tumor reduction, and that the timing of tumor samples for assessment of response is important. Although inherent complete resistance of breast cancer to paclitaxel occurs in a minority of patients, mechanisms of acquired or partial resistance require further study. However, the initial apoptotic response to paclitaxel has been shown to transiently reduce both cell density and intratumoral pressure, providing a window of time when there can be improved penetration of paclitaxel into the tumor. Thus a strong initial apoptotic response can set up a compounding benefit from subsequent treatments. Knowledge of breast cancer response to paclitaxel in vivo could lead to therapeutic strategies that enhance the apoptotic response and optimize the dosing schedule, to improve the tumor reduction for most patients.

Fine-needle aspiration of pancreatic extramedullary plasmacytoma: possible confusion with islet cell tumor.

We report two cases of extramedullary plasmacytoma (EMP) of the pancreas, diagnosed by fine-needle aspiration (FNA). Plasmacytoma of the pancreas is a rare neoplasm with only 12 cases recorded in the literature. Because of its scarcity and the cytomorphologic similarity between plasma cells and endocrine cells, EMP of the pancreas may be confused with neuroendocrine (islet cell) tumors of the pancreas. Immunohistochemical staining for light chain and/or neuroendocrine markers will prevent diagnostic error when interpreting plasmacytoid neoplasms of any site susceptible to endocrine tumors, including the pancreas.

Detection of the apoptosis-suppressing oncoprotein bc1-2 in hormone-refractory human prostate cancers.

The oncoprotein encoded by bc1-2 is unique because of its intracellular location (a mitochondrial membrane protein) and apparent mode of action (suppression of apoptosis). To date, this oncogene has been associated only with the development of certain forms of human B-cell lymphoma. In this report, we describe our experience with a monoclonal antibody made against a synthetic peptide for bc1-2 that can recognize the bc1-2 protein and identify cells in human prostate glands expressing this proto-oncogene with in situ immunohistochemical procedures. These procedures were utilized to survey a series of 62 human tissues to evaluate whether bc1-2 might have a role in the developing prostate gland or in prostate oncogenesis. While all primordial epithelial cells in a fetal prostate gland immunostain for bc1-2, normal and hypertrophic prostate glands of the adult show bc1-2 expression restricted to the basal cells. All epithelial cells in areas of prostatic intraepithelial neoplasia were stained by this antibody, as were most (62%) localized invasive prostatic carcinomas. In contrast, all primary prostatic carcinomas and metastases obtained from metastatic prostate cancer patients after hormone treatment (hormone-refractory tumors) stained positive for bc1-2. This study demonstrates that the oncoprotein encoded by bc1-2 can be detected at sequential stages in the natural history of human prostate cancer. Since the bc1-2 oncoprotein is known to suppress the cellular response to apoptotic stimuli, it will be important to determine whether bc1-2 expression is a factor in the development of prostate cancers and in the survival of hormone-refractory prostate cancer cells.

Correlation of cervical cytology and human papillomavirus DNA detection in postmenopausal women.

This study correlated the cytological features of cervical swabs from postmenopausal women with the histological findings and the detection of human papillomavirus (HPV) using filter hybridization analysis. Of 17 postmenopausal women seen at colposcopy for an abnormal Papanicolaou (Pap) smear, most often squamous atypia, HPV DNA was detected in one (6%) cervical swab. Biopsy-proven cervical squamous intraepithelial lesions (SILs) were noted in two of the 17 (12%). The HPV DNA detection rate was equivalent to that found in 47 postmenopausal women who had hysterectomies for noncervical disease. The rates of HPV detection and biopsy-proven SIL in premenopausal women seen at colposcopy was 55% and 66%, respectively. The HPV DNA detection rate for postmenopausal women with biopsy-proven SILs as determined by in situ hybridization was 19/26 (73%), which is equivalent to the rate in premenopausal women. It is concluded that squamous atypia in postmenopausal women is rarely associated with either biopsy-proven SIL or HPV DNA detection and thus, in many cases, may represent atrophic changes.

Immunohistochemistry and proliferative activity in Lhermitte-Duclos disease.

We have evaluated a recurrence of Lhermitte-Duclos disease by immunohistochemistry for Purkinje cell markers and proliferative activity (proliferating cell nuclear antigen), by electron microscopy and for DNA ploidy (image analysis). While most of the abnormal neurons in the lesion appear to be derived from granule cells, several Purkinje cell specific polyclonal and monoclonal antibodies, including L7, PEP 19 and calbindin, labeled a minor subpopulation. Staining with monoclonal antibodies to proliferating cell nuclear antigen and measuring cell DNA index and ploidy with a cell image analyzer revealed no proliferative activity. Electron microscopy findings were similar to those previously reported. In spite of its recurrence, our findings support the notion that Lhermitte-Duclos disease is malformative, not neoplastic, and that the characteristic neurons are derived predominantly but not exclusively from a non-Purkinje cell source, probably the granule cell.

Human papillomavirus detection in adenocarcinoma of the anus.

Human papillomavirus (HPV) is associated with carcinoma of the cervix but not with carcinoma of the endometrium. HPV 16 is the type most commonly detected in squamous cell carcinomas from this site, whereas HPV 18 predominates in adenocarcinomas. We analyzed eight anal carcinomas for HPV DNA using the polymerase chain reaction and type-specific (open reading frame E6) primers for HPV 16, 18, 31, and 35. HPV DNA sequences were amplified in two of six anal adenocarcinomas and, in each case, the type was HPV 18. Sequences homologous to HPV 16 were amplified in each of two anal squamous cell carcinomas; one also contained HPV 18. No amplification was detected in any of seven adenocarcinomas of the rectum or colon or in three adenomatous polyps of the colon. It is concluded that HPV is associated with anal adenocarcinomas but not colorectal adenocarcinomas. The reason(s) why HPV is associated with adenocarcinoma of the anus and cervix but not with the rectum and endometrium, despite the close proximity, requires further study.