Prognostic prediction value of the clinical-radiomics tumour-stroma ratio in locally advanced rectal cancer.

PURPOSE: To develop and validate a radiomics model based on high-resolution T2WI and a clinical-radiomics model for tumour-stroma ratio (TSR) evaluation with a gold standard of TSR evaluated by rectal specimens without therapeutic interference and further apply them in prognosis prediction of locally advanced rectal cancer (LARC) patients who received neoadjuvant chemoradiotherapy. METHODS: A total of 178 patients (mean age: 59.35, range 20-85 years; 65 women and 113 men) with rectal cancer who received surgery alone from January 2016 to October 2020 were enrolled and randomly separated at a ratio of 7:3 into training and validation sets. A senior radiologist reviewed after 2 readers manually delineated the whole tumour in consensus on preoperative high-resolution T2WI in the training set. A total of 1046 features were then extracted, and recursive feature elimination embedded with leave-one-out cross validation was applied to select features, with which an MR-TSR evaluation model was built containing 6 filtered features via a support vector machine classifier trained by comparing patients' pathological TSR. Stepwise logistic regression was employed to integrate clinical factors with the radiomics model (Fusion-TSR) in the training set. Later, the MR-TSR and Fusion-TSR models were replicated in the validation set for diagnostic effectiveness evaluation. Subsequently, 243 patients (mean age: 53.74, range 23-74 years; 63 women and 180 men) with LARC from October 2012 to September 2017 who were treated with NCRT prior to surgery and underwent standard pretreatment rectal MR examination were enrolled. The MR-TSR and Fusion-TSR were applied, and the Kaplan-Meier method and log-rank test were used to compare the survival of patients with different MR-TSR and Fusion-TSR. Cox proportional hazards regression was used to calculate the hazard ratio (HR). RESULTS: Both the MR-TSR and Fusion-TSR models were validated with favourable diagnostic power: the AUC of the MR-TSR was 0.77 (p = 0.01; accuracy = 69.8 %, sensitivity = 88.9 %, specificity = 65.9 %, PPV = 34.8 %, NPV = 96.7 %), while the AUC of the Fusion-TSR was 0.76 (p = 0.014; accuracy = 67.9 %, sensitivity = 88.9 %, specificity = 63.6 %, PPV = 33.3 %, NPV = 96.6 %), outperforming their effectiveness in the training set: the AUC of the MR-TSR was 0.65 (p = 0.035; accuracy = 66.4 %, sensitivity = 61.9 %, specificity = 67.3 %, PPV = 27.7 %, NPV = 90.0 %), while the AUC of the Fusion-TSR was 0.73 (p = 0.001; accuracy = 73.6 %, sensitivity = 71.4 %, specificity = 74.0 %, PPV = 35.73 %, NPV = 92.8 %). With further prognostic analysis, the MR-TSR was validated as a significant prognostic factor for DFS in LARC patients treated with NCRT (p = 0.020, HR = 1.662, 95 % CI = 1.077-2.565), while the Fusion-TSR was a significant prognostic factor for OS (p = 0.005, HR = 2.373, 95 % CI = 1.281-4.396). CONCLUSIONS: We developed and validated a radiomics TSR and a clinical-radiomics TSR model and successfully applied them to better risk stratification for LARC patients receiving NCRT and for better decision making.

Node-negative colon cancer: histological, molecular, and stromal features predicting disease recurrence.

BACKGROUND: Within the group of node-negative colon cancer patients, presumed to have a good prognosis, a significant percentage of patients develops cancer-recurrence. Current high-risk features prove inadequate to select these particular high-risk patients. In the process of tailor-made care and shared decision-making the need to identify these patients grows. In this study we investigate the value of adding molecular markers and the tumour-stroma ratio (TSR) to conventional histological tumour staging methods to improve the selection of high risk patients. METHODS: We retrospectively analysed 201 patients diagnosed with TNM-stage I-II colon cancer and treated by complete oncological resection between November 1st 2002 and December 31st 2012 at the Jeroen Bosch Hospital. Conventional histological tumour staging, BRAF mutations, KRAS mutations, MSI status and TSR were determined. Differences between groups based on TSR and mutation status, in disease free survival were analysed using Cox-Regression analyses. RESULTS: Poorly differentiated histology (p = 0.002), high-TSR (p = 0.033), BRAF-mutation (p = 0.008) and MSI (p = 0.011) were identified as significant risk factors for cancer recurrence. The risk of recurrence increased in the presence of both a BRAF-mutation and high-TSR compared to the absence of both factors or presence of only one factor (HR = 3.66 BRAF-mt/TSR-low (p = 0.006), HR 2.82 BRAF-wt/TSR-high (p = 0.015), HR = 4.39 BRAF-mt/TSR-high (p = 0.023)). This was also seen in tumours with MSI and high-TSR (HR = 2.46 MSS/TSR-high (p = 0.041), HR = 3.31 MSI/TSR-high (p = 0.045). CONCLUSION: Judging by the higher HR for the combination of the prognostic factors TSR and BRAF compared to the HRs of these prognostic factors individually, the prognostication for disease free survival can be improved by determining both TSR and BRAF instead of BRAF alone, as is done in current daily practise. In this study MSI also shows additional value to TSR in the prognostication of disease free survival. Adopting TSR into daily diagnostics will be of additional value next to currently used molecular markers in risk stratification of patients with node negative colon cancer and is therefore advised.

Tumour-stroma ratio (TSR) in breast cancer: comparison of scoring core biopsies versus resection specimens.

PURPOSE: Tumour-stroma ratio (TSR) is an important prognostic and predictive factor in several tumour types. The aim of this study is to determine whether TSR evaluated in breast cancer core biopsies is representative of the whole tumour. METHOD: Different TSR scoring methods, their reproducibility, and the association of TSR with clinicopathological characteristics were investigated in 178 breast carcinoma core biopsies and corresponding resection specimens. TSR was assessed by two trained scientists on the most representative H&E-stained digitised slides. Patients were treated primarily with surgery between 2010 and 2021 at Semmelweis University, Budapest. RESULTS: Ninety-one percent of the tumours were hormone receptor (HR)-positive (luminal-like). Interobserver agreement was highest using 100 × magnification (κcore = 0.906, κresection specimen = 0.882). The agreement between TSR of core biopsies and resection specimens of the same patients was moderate (κ = 0.514). Differences between the two types of samples were most frequent in cases with TSR scores close to the 50% cut-off point. TSR was strongly correlated with age at diagnosis, pT category, histological type, histological grade, and surrogate molecular subtype. A tendency was identified for more recurrences among stroma-high (SH) tumours (p = 0.07). Significant correlation was detected between the TSR and tumour recurrence in grade 1 HR-positive breast cancer cases (p = 0.03). CONCLUSIONS: TSR is easy to determine and reproducible on both core biopsies and in resection specimens and is associated with several clinicopathological characteristics of breast cancer. TSR scored on core biopsies is moderately representative for the whole tumour.

Tumour-stroma ratio is a valuable prognostic factor for oral tongue squamous cell carcinoma.

OBJECTIVE: The objectives of this study were to estimate the prognostic value of the tumour-stroma ratio (TSR) and tumour budding (TB) in oral tongue squamous cell carcinoma (OTSCC) and to establish a reliable model to predict the outcome of OTSCC patients. METHODS: A total of 103 patients surgically treated at our hospital were enrolled in this study. Chi-square tests, Kaplan-Meier analyses and Cox proportional hazards regression models were performed for statistical analysis. RESULTS: Fifty-six patients were categorized as stroma-rich, and 47 patients were categorized as stroma-poor. Only pathological grade was associated with the TSR (p = 0.017). Kaplan-Meier analysis showed that stroma-rich, high-intensity budding and high risk groups were associated with worse prognosis. The Cox regression model showed that the TSR was an independent risk factor for OTSCC patients prognosis, and the high risk group was also related to poor prognosis (p < 0.05). TB was significantly associated with poor prognosis but was not an independent risk factor. CONCLUSIONS: We found that patients in the stroma-rich group had a worse long-term prognosis. The TSR is an independent risk factor for OTSCC patients' outcome. In addition, a risk model that combined the TSR and TB proved to be valuable for predicting OTSCC patients' outcome.

Tumour-stroma ratio to predict pathological response to neo-adjuvant treatment in rectal cancer.

INTRODUCTION: Management of rectal cancer has advanced, with an increasing use of neoadjuvant chemoradiotherapy (nCRT). This opens options for organ preserving treatment for those with a major response to nCRT. However, the degree of clinical response, based on MRI and post-treatment biopsies, only poorly matches the degree of actual pathological response. In order to select patients with major pathological response without surgical resection, it is of importance to define tumour markers predicting the degree of pathological response to nCRT. The intra-tumoural tumour-stroma ratio (TSR) might be this marker. METHODS: TSR in pre-treatment biopsies was estimated according to the method described by van Pelt et al. The degree of pathological response was assessed on the tumour resection according to tumour regression grading (TRG) by Mandard. The primary endpoint of this study was the difference in pathological response to nCRT between TSR-high and TSR-low groups. RESULTS: We found that 26.2% of patients with major response was classified as TSR-high, while 73.8% of patients were classified as TSR-low. A high TSR in pre-treatment biopsies was associated with a lower chance of major-response to nCRT (OR = 0.37, 95%CI; 0.19-0.73), p = 0.004), independent of tumour stage and time between nCRT and surgery. CONCLUSION: In rectal cancer, TSR in pre-treatment biopsies predicts pathologic response to nCRT, with a high TSR bringing twice the risk of poor to no response compared to low TSR. In future, assessment of TSR may fulfil a role in a therapeutic algorithm identifying patients who will or will not respond to nCRT prior to treatment initiation.

Attention-guided sampling for colorectal cancer analysis with digital pathology.

Improvements to patient care through the development of automated image analysis in pathology are restricted by the small image patch size that can be processed by convolutional neural networks (CNNs), when compared to the whole-slide image (WSI). Tile-by-tile processing across the entire WSI is slow and inefficient. While this may improve with future computing power, the technique remains vulnerable to noise from uninformative image areas. We propose a novel attention-inspired algorithm that selects image patches from informative parts of the WSI, first using a sparse randomised grid pattern, then iteratively re-sampling at higher density in regions where a CNN classifies patches as tumour. Subsequent uniform sampling across the enclosing region of interest (ROI) is used to mitigate sampling bias. Benchmarking tests informed the adoption of VGG19 as the main CNN architecture, with 79% classification accuracy. A further CNN was trained to separate false-positive normal epithelium from tumour epithelium, in a novel adaptation of a two-stage model used in brain imaging. These subsystems were combined in a processing pipeline to generate spatial distributions of classified patches from unseen WSIs. The ROI was predicted with a mean F1 (Dice) score of 86.6% over 100 evaluation WSIs. Several algorithms for evaluating tumour-stroma ratio (TSR) within the ROI were compared, giving a lowest root mean square (RMS) error of 11.3% relative to pathologists' annotations, against 13.5% for an equivalent tile-by-tile pipeline. Our pipeline processed WSIs between 3.3x and 6.3x faster than tile-by-tile processing. We propose our attention-based sampling pipeline as a useful tool for pathology researchers, with the further potential for incorporating additional diagnostic calculations.

Characteristics of tumour stroma in regional lymph node metastases in colorectal cancer patients: a theoretical framework for future diagnostic imaging with FAPI PET/CT

PurposeThe recently developed fibroblast activation protein inhibitor (FAPI) tracer for PET/CT, binding tumour-stromal cancer-associated fibroblasts, is a promising tool for detection of positive lymph nodes. This study provides an overview of features, including sizes and tumour-stromal content, of lymph nodes and their respective lymph node metastases (LNM) in colorectal cancer (CRC), since literature lacks on whether LNMs contain sufficient stroma to potentially allow FAPI-based tumour detection.MethodsHaematoxylin and eosin-stained tissue slides from 73 stage III colon cancer patients were included. Diameters and areas of all lymph nodes and their LNMs were assessed, the amount of stroma by measuring the stromal compartment area, the conventional and total tumour-stroma ratios (TSR-c and TSR-t, respectively), as well as correlations between these parameters. Also, subgroup analysis using a minimal diameter cut off of 5.0 mm was performed.ResultsIn total, 126 lymph nodes were analysed. Although positive correlations were observed between node and LNM for diameter and area (r = 0.852, p < 0.001 and r = 0.960, p < 0.001, respectively), and also between the LNM stromal compartment area and nodal diameter (r = 0.612, p < 0.001), nodal area (r = 0.747, p < 0.001) and LNM area (r = 0.746, p < 0.001), novel insight was that nearly all (98%) LNMs contained stroma, with median TSR-c scores of 35% (IQR 20–60%) and TSR-t of 20% (IQR 10–30%). Moreover, a total of 32 (25%) positive lymph nodes had a diameter of < 5.0 mm.ConclusionIn LNMs, stroma is abundantly present, independent of size, suggesting a role for FAPI PET/CT in improved lymph node detection in CRC. (AU)

A high tumour-stroma ratio (TSR) in colon tumours and its metastatic lymph nodes predicts poor cancer-free survival and chemo resistance

PurposeDespite known high-risk features, accurate identification of patients at high risk of cancer recurrence in colon cancer remains a challenge. As tumour stroma plays an important role in tumour invasion and metastasis, the easy, low-cost and highly reproducible tumour-stroma ratio (TSR) could be a valuable prognostic marker, which is also believed to predict chemo resistance.MethodsTwo independent series of patients with colon cancer were selected. TSR was estimated by microscopic analysis of 4 µm haematoxylin and eosin (H&E) stained tissue sections of the primary tumour and the corresponding metastatic lymph nodes. Patients were categorized as TSR-low (≤ 50%) or TSR-high (> 50%). Differences in overall survival and cancer-free survival were analysed by Kaplan–Meier curves and cox-regression analyses. Analyses were conducted for TNM-stage I–II, TNM-stage III and patients with an indication for chemotherapy separately.ResultsWe found that high TSR was associated with poor cancer-free survival in TNM-stage I–II colon cancer in two independent series, independent of other known high-risk features. This association was also found in TNM-stage III tumours, with an additional prognostic value of TSR in lymph node metastasis to TSR in the primary tumour alone. In addition, high TSR was found to predict chemo resistance in patients receiving adjuvant chemotherapy after surgical resection of a TNM-stage II–III colon tumour.ConclusionIn colon cancer, the TSR of both primary tumour and lymph node metastasis adds significant prognostic value to current pathologic and clinical features used for the identification of patients at high risk of cancer recurrence, and also predicts chemo resistance.

Characteristics of tumour stroma in regional lymph node metastases in colorectal cancer patients: a theoretical framework for future diagnostic imaging with FAPI PET/CT.

PURPOSE: The recently developed fibroblast activation protein inhibitor (FAPI) tracer for PET/CT, binding tumour-stromal cancer-associated fibroblasts, is a promising tool for detection of positive lymph nodes. This study provides an overview of features, including sizes and tumour-stromal content, of lymph nodes and their respective lymph node metastases (LNM) in colorectal cancer (CRC), since literature lacks on whether LNMs contain sufficient stroma to potentially allow FAPI-based tumour detection. METHODS: Haematoxylin and eosin-stained tissue slides from 73 stage III colon cancer patients were included. Diameters and areas of all lymph nodes and their LNMs were assessed, the amount of stroma by measuring the stromal compartment area, the conventional and total tumour-stroma ratios (TSR-c and TSR-t, respectively), as well as correlations between these parameters. Also, subgroup analysis using a minimal diameter cut off of 5.0 mm was performed. RESULTS: In total, 126 lymph nodes were analysed. Although positive correlations were observed between node and LNM for diameter and area (r = 0.852, p < 0.001 and r = 0.960, p < 0.001, respectively), and also between the LNM stromal compartment area and nodal diameter (r = 0.612, p < 0.001), nodal area (r = 0.747, p < 0.001) and LNM area (r = 0.746, p < 0.001), novel insight was that nearly all (98%) LNMs contained stroma, with median TSR-c scores of 35% (IQR 20-60%) and TSR-t of 20% (IQR 10-30%). Moreover, a total of 32 (25%) positive lymph nodes had a diameter of < 5.0 mm. CONCLUSION: In LNMs, stroma is abundantly present, independent of size, suggesting a role for FAPI PET/CT in improved lymph node detection in CRC.

A high tumour-stroma ratio (TSR) in colon tumours and its metastatic lymph nodes predicts poor cancer-free survival and chemo resistance.

PURPOSE: Despite known high-risk features, accurate identification of patients at high risk of cancer recurrence in colon cancer remains a challenge. As tumour stroma plays an important role in tumour invasion and metastasis, the easy, low-cost and highly reproducible tumour-stroma ratio (TSR) could be a valuable prognostic marker, which is also believed to predict chemo resistance. METHODS: Two independent series of patients with colon cancer were selected. TSR was estimated by microscopic analysis of 4 µm haematoxylin and eosin (H&E) stained tissue sections of the primary tumour and the corresponding metastatic lymph nodes. Patients were categorized as TSR-low (≤ 50%) or TSR-high (> 50%). Differences in overall survival and cancer-free survival were analysed by Kaplan-Meier curves and cox-regression analyses. Analyses were conducted for TNM-stage I-II, TNM-stage III and patients with an indication for chemotherapy separately. RESULTS: We found that high TSR was associated with poor cancer-free survival in TNM-stage I-II colon cancer in two independent series, independent of other known high-risk features. This association was also found in TNM-stage III tumours, with an additional prognostic value of TSR in lymph node metastasis to TSR in the primary tumour alone. In addition, high TSR was found to predict chemo resistance in patients receiving adjuvant chemotherapy after surgical resection of a TNM-stage II-III colon tumour. CONCLUSION: In colon cancer, the TSR of both primary tumour and lymph node metastasis adds significant prognostic value to current pathologic and clinical features used for the identification of patients at high risk of cancer recurrence, and also predicts chemo resistance.

Tumour-stroma ratio outperforms tumour budding as biomarker in colon cancer: a cohort study.

The tumour-stroma ratio (TSR) and tumour budding (TB) are two high-risk factors with potential to be implemented in the next TNM classification. The aim of the current study was to evaluate the practical application of the two biomarkers based on reproducibility, independency and prognostic value. Patients diagnosed with stage II or III colon cancer who underwent surgery between 2005 and 2016 were included. Both TSR and TB were scored on haematoxylin and eosin-stained tissue sections. The TSR, based on the relative amount of stroma, was scored in increments of 10%. TB was scored following the consensus guidelines; a bud was defined as ≤ 4 tumour cells. For analysis, three categories were used. Cohen's kappa was used for reproducibility. The prognostic value was determined with survival analysis. In total, 246 patients were included. The TSR distribution was N = 137 (56%) stroma-low and N = 109 (44%) stroma-high. The TB distribution was TB-low N = 194 (79%), TB-intermediate N = 35 (14%) and TB-high N = 17 (7%). The reproducibility of the TSR was good (interobserver agreement kappa = 0.83 and intraobserver agreement kappa = 0.82), whereas the inter- and intraobserver agreement for scoring TB was moderate (kappa 0.47 and 0.45, respectively). The survival analysis showed an independent prognostic value for disease-free survival for TSR (HR 1.57; 95% CI 1.01-2.44; p = 0.048) and for TB-high (HR 2.01; 95% CI 1.02-3.96; p = 0.043). Based on current results, we suggest the TSR is a more reliable parameter in daily practice due to better reproducibility and independent prognostic value for disease-free survival.

The value of the tumour-stroma ratio for predicting neoadjuvant chemoradiotherapy response in locally advanced rectal cancer: a case control study.

BACKGROUND: The tumour-stroma ratio (TSR) is recognized as a practical prognostic factor in colorectal cancer. However, TSR assessment generally utilizes surgical specimens. This study aims to investigate whether the TSR evaluated from preoperative biopsy specimens by a semi-automatic quantification method can predict the response after neoadjuvant chemoradiotherapy (nCRT) of patients with locally advanced rectal cancer (LARC). METHODS: A total of 248 consecutive patients diagnosed with LARC and treated with nCRT followed by resection were included. Haematoxylin and eosin (HE)-stained sections of biopsy specimens were collected, and the TSR was evaluated by a semi-automatic quantification method and was divided into three categories, using the cut-offs determined in the whole cohort to balance the proportion of patients in each category. The response to nCRT was evaluated on the primary tumour resection specimen by an expert pathologist using the four-tier tumour regression grade (TRG) system. RESULTS: The TSR can discriminate patients that are major-responders (TRG 0-1) from patients that are non-responders (TRG 2-3). Patients were divided into stroma-low (33.5%), stroma-intermediate (33.9%), and stroma-high (32.7%) groups using 56.3 and 72.8% as the cutoffs. In the stroma-low group, 58 (69.9%) patients were major-responders, and only 39 (48.1%) patients were considered major-responders in the stroma-high group (P = 0.018). Multivariate analysis showed that the TSR was the only pre-treatment predictor of response to nCRT (adjusted odds ratio 0.40, 95% confidence interval 0.21-0.76, P = 0.002). CONCLUSION: An elevated TSR in preoperative biopsy specimens is an independent predictor of nCRT response in LARC. This semi-automatic quantified TSR could be easily translated into routine pathologic assessment due to its reproducibility and reliability.

Tumour-stroma ratio has poor prognostic value in nonpedunculated T1 colorectal cancer: A multicentre case-cohort study.

BACKGROUND: Current risk stratification models for early invasive (T1) colorectal cancer are not able to discriminate accurately between prognostic favourable and unfavourable tumours, resulting in over-treatment of a large (>80%) proportion of T1 colorectal cancer patients. The tumour-stroma ratio (TSR), which is a measure for the relative amount of desmoplastic tumour stroma, is reported to be a strong independent prognostic factor in advanced-stage colorectal cancer, with a high stromal content being associated with worse prognosis and survival. We aimed to investigate whether the TSR predicts clinical outcome in patients with non-pedunculated T1 colorectal cancer. METHODS: Haematoxylin and eosin (H&E)-stained tumour tissue slides from a retrospective multicentre case cohort of patients with nonpedunculated surgically treated T1 colorectal cancer were assessed for TSR by two independent observers who were blinded for clinical outcomes. The primary end point was adverse outcome, which was defined as the presence of lymph node metastasis in the resection specimen or colorectal cancer recurrence during follow-up. RESULTS: All 261 patients in the case cohort had H&E slides available for TSR scoring. Of these, 183 were scored as stroma-low, and 78 were scored as stroma-high. There was moderate inter-observer agreement κ = 0.42). In total, 41 patients had lymph node metastasis, 17 patients had recurrent cancer and five had both. Stroma-high tumours were not associated with an increased risk for an adverse outcome (adjusted hazard ratio = 0.66, 95% confidence interval 0.37-1.18; p = 0.163). CONCLUSIONS: Our study emphasises that existing prognosticators may not be simply extrapolated to T1 colorectal cancers, even though their prognostic value has been widely validated in more advanced-stage tumours.

Tumour-stroma ratio has poor prognostic value in non-pedunculated T1 colorectal cancer: A multi-centre case-cohort study.

BACKGROUND: Current risk stratification models for early invasive (T1) colorectal cancer are not able to discriminate accurately between prognostic favourable and unfavourable tumours, resulting in over-treatment of a large (>80%) proportion of T1 colorectal cancer patients. The tumour-stroma ratio (TSR), which is a measure for the relative amount of desmoplastic tumour stroma, is reported to be a strong independent prognostic factor in advanced-stage colorectal cancer, with a high stromal content being associated with worse prognosis and survival. We aimed to investigate whether the TSR predicts clinical outcome in patients with non-pedunculated T1 colorectal cancer. METHODS: Hematoxylin and eosin (H&E)-stained tumour tissue slides from a retrospective multi-centre case cohort of patients with non-pedunculated surgically treated T1 colorectal cancer were assessed for TSR by two independent observers who were blinded for clinical outcomes. The primary end point was adverse outcome, which was defined as the presence of lymph node metastasis in the resection specimen or colorectal cancer recurrence during follow-up. RESULTS: All 261 patients in the case cohort had H&E slides available for TSR scoring. Of these, 183 were scored as stroma-low, and 78 were scored as stroma-high. There was moderate inter-observer agreement (κ = 0.42). In total, 41 patients had lymph node metastasis, 17 patients had recurrent cancer and five had both. Stroma-high tumours were not associated with an increased risk for an adverse outcome (adjusted hazard ratio = 0.66, 95% confidence interval 0.37-1.18; p = 0.163). CONCLUSIONS: Our study emphasises that existing prognosticators may not be simply extrapolated to T1 colorectal cancers, even though their prognostic value has been widely validated in more advanced-stage tumours.

Artificial intelligence quantified tumour-stroma ratio is an independent predictor for overall survival in resectable colorectal cancer.

BACKGROUND: An artificial intelligence method could accelerate the clinical implementation of tumour-stroma ratio (TSR), which has prognostic relevance in colorectal cancer (CRC). We, therefore, developed a deep learning model for the fully automated TSR quantification on routine haematoxylin and eosin (HE) stained whole-slide images (WSI) and further investigated its prognostic validity for patient stratification. METHODS: We trained a convolutional neural network (CNN) model using transfer learning, with its nine-class tissue classification performance evaluated in two independent test sets. Patch-level segmentation on WSI HE slides was performed using the model, with TSR subsequently derived. A discovery (N=499) and validation cohort (N=315) were used to evaluate the prognostic value of TSR for overall survival (OS). FINDINGS: The CNN-quantified TSR was a prognostic factor, independently of other clinicopathologic characteristics, with stroma-high associated with reduced OS in the discovery (HR 1.72, 95% CI 1.24-2.37, P=0.001) and validation cohort (2.08, 1.26-3.42, 0.004). Integrating TSR into a Cox model with other risk factors showed improved prognostic capability. INTERPRETATION: We developed a deep learning model to quantify TSR based on histologic WSI of CRC and demonstrated its prognostic validity for patient stratification for OS in two independent CRC patient cohorts. This fully automatic approach allows for the objective and standardised application while reducing pathologists' workload. Thus, it can potentially be of significant aid in clinical prognosis prediction and decision-making. FUNDING: National Key Research and Development Program of China, National Science Fund for Distinguished Young Scholar, and National Science Foundation for Young Scientists of China.

Prognostication for oral squamous cell carcinoma patients based on the tumour-stroma ratio and tumour budding.

AIMS: Previous studies have demonstrated that the tumour-stroma ratio (TSR) and tumour budding are of prognostic value for oral squamous cell carcinomas (OSCCs). The aim of this study was to evaluate the prognostic significance of those histological parameters, individually and in combination, for OSCC. METHODS AND RESULTS: The TSR and tumour budding (the presence of five or more buds at the invasive front) were estimated in 254 patients with OSCC. The clinicopathological association was investigated with a chi-square test, and the prognostic significance (cancer-specific survival and disease-free survival) was verified with Kaplan-Meier analysis and the Cox proportional hazard model. The TSR (≥50%, stroma-rich) was significantly and independently associated with both shortened cancer-specific survival and poor disease-free survival, whereas tumour budding was significantly associated with reduced cancer-specific survival. The TSR/tumour budding model was independently associated with a high risk of cancer mortality and recurrence (disease-free survival). In patients with early-stage tumours (clinical stage I and II, n = 103), the TSR, tumour budding and the TSR/tumour budding model were significantly associated with both cancer-related death and recurrence, whereas, in advanced-stage tumours (clinical stage III and IV, n = 144), only the TSR and the TSR/tumour budding model were significantly associated with cancer-specific survival. CONCLUSIONS: The TSR, tumour budding and their combination provide significant information on OSCC outcome, suggesting that their incorporation in the routine evaluation of histopathological specimens might be useful in prognostication for OSCC patients.

The intra-tumoural stroma in patients with breast cancer increases with age.

PURPOSE: The tumour microenvironment in older patients is subject to changes. The tumour-stroma ratio (TSR) was evaluated in order to estimate the amount of intra-tumoural stroma and to evaluate the prognostic value of the TSR in older patients with breast cancer (≥ 70 years). METHODS: Two retrospective cohorts, the FOCUS study (N = 619) and the Nottingham Breast Cancer series (N = 1793), were used for assessment of the TSR on haematoxylin and eosin stained tissue slides. RESULTS: The intra-tumoural stroma increases with age in the FOCUS study and the Nottingham Breast Cancer series (B 0.031, 95% CI 0.006-0.057, p = 0.016 and B 0.034, 95% CI 0.015-0.054, p < 0.001, respectively). Fifty-one per cent of the patients from the Nottingham Breast Cancer series < 40 years had a stroma-high tumour compared to 73% of the patients of ≥ 90 years from the FOCUS study. The TSR did not validate as an independent prognostic parameter in patients ≥ 70 years. CONCLUSIONS: The intra-tumoural stroma increases with age. This might be the result of an activated tumour microenvironment. The TSR did not validate as an independent prognostic parameter in patients ≥ 70 years in contrast to young women with breast cancer as published previously.

Tumour Budding and Tumour Stroma Ratio are Reliable Predictors for Death and Recurrence in Elderly Stage I Colon Cancer Patients.

AIM: Tumour budding (BD) and Tumour stroma ratio (TSR) are considered valuable survival parameters for colon cancer (CC), but it is still unclear whether these parameters predict a poor prognosis. This study aimed to determine the survival effect of TB and TSR in elderly stage I CC patients. METHODS: We evaluated these parameters in eighty-eight stage I CC patients who underwent surgical treatment alone between 1998 and 2015. The relationship between these parameters and age was investigated. Model A was used for methodology. RESULTS: In elderly patients (≥68), these parameters obtained more successful results for following analysis: relationship between prognostic factors [lymphatic invasion (BD, p = 0.003; TSR, p = 0.003), perineural invasion (BD, p = 0.016; TSR, p = 0.004), tumour deposites (BD, p = 0.005; TSR, p = 0.033), MSI (BD, p=0.031; TSR, p=0.012), etc.], correlation of estimates (BD, r=0.724; TSR, r=0.703), and reproducibility of study (BD [Kappa (ĸ)]=0.53-0.75; TSR (ĸ) =0.56-0.71). Also, the cut-off values were useful for BD (area of under ROC (AUC)=0.816 [0.707-0.925]) and TSR (AUC=0.810 [0.697-0.924]). In univariate analysis, in elderly patients, these parameters had a poor 5-year survival for RFS (p<0.001[BD], p=0.001[TSR]), OS (p<0.001[BD], p=0.005[TSR]), and LR (p=0.008[BD], p=0.034[TSR]). Multivariate analysis confirmed that these parameters are independent worse predictors for RFS (BD: 1.42[HR], p=0.002; TSR: 1.50[HR], p=0.001), OS (BD: 1.38[HR], p=0.014; TSR: 1.42[HR], p=0.005), and LR (BD: 1.61[HR], p=0.034) in elderly patients. CONCLUSIONS: Our results confirm that BD and TSR are reliable indexes for poor survival in elderly stage I CC patients. We recommend using model A for successful results and standardization.

Prognostic value of low CDX2 expression in colorectal cancers with a high stromal content - a short report.

PURPOSE: Lack of expression of the intestinal transcription factor CDX2 in colorectal cancer (CRC) identifies patients with a poor prognosis. This biomarker has previously been suggested to be prognostic in CRCs with a high stromal content based on mRNA expression data. We investigated the prognostic value of CDX2 expression in microsatellite stable CRC stratified by stromal content using microscopy-based techniques. METHODS AND RESULTS: The study included a cohort of 236 patients with stage I-IV CRC. We assessed by microscopy the tumour-stroma ratio (TSR) and the immunohistochemical CDX2 intensity. We found that patients of the stroma-high group had a worse prognosis compared to those of the stroma-low group [disease-free survival in a multivariate analysis (DFSmultivariate) HR 1.52 (95% CI 1.05-2.21)]. In our cohort, low CDX2 expression (14.6%) showed prognostic value for DFSmultivariate [HR 1.93 (95% CI 1.16-3.23)]. Interestingly, when stratifying the cohort by TSR, no prognostic difference was observed related to CDX2 expression in stroma-low tumours. However, CDX2 expression was found to be prognostic within the stroma-high group [DFSmultivariate HR 3.02 (95% CI 1.49-6.13)]. The p value for interaction between TSR and CDX2 status was borderline significant in DFS (p = 0.071). CONCLUSIONS: The present study confirms a poor outcome of patients with stroma-high tumours. Low CDX2 expression in tumours with a high stromal content identified patients with a particularly poor prognosis. The present study did not reveal a clear difference in TSR associated with CDX2 status and survival. This method, solely based on microscopy, identifies patients who have a high risk of relapse and a poor outcome, and who may benefit from targeted therapy.

The prognostic value of tumour-stroma ratio in primary breast cancer with special attention to triple-negative tumours: a review.

PURPOSE: There is a strong need to improve the prognostication of breast cancer patients in order to prevent over- and undertreatment, especially when considering adjuvant chemotherapy. Tumour stroma characteristics might be valuable in predicting disease progression. METHODS: Studies regarding the prognostic value of tumour-stroma ratio (TSR) in breast cancer are evaluated. RESULTS: A high stromal content is related to a relatively poor prognosis. The most pronounced prognostic effect of this parameter seems to be observed in the triple-negative breast cancer (TNBC) subtype. CONCLUSIONS: TSR assessment might represent a simple, fast and reproducible prognostic factor at no extra costs, and could possibly be incorporated into routine pathological diagnostics. Despite these advantages, a robust clinical validation of this parameter has yet to be established in prospective studies.