Predict Tumor Growth Research Articles

Multiscale in silico knowledge-based model predicts tumor growth in advanced EGFR+ NSCLC patients on gefitinib.

e20565 Background: Tumor heterogeneity and treatment resistance remains a continued threat to patients with advanced non-small cell lung cancer (NSCLC), despite the emergence of targeted medicines, e.g. against epidermal growth factor receptor (EGFR) alterations. We developed an in silico EGFR mutated lung adenocarcinoma (EGFR+ LUAD) model to predict the effect of known oncogenic EGFR mutations (common EGFR mutations, EGFR exon 20 insertions). The model provides mechanistic representation of tumor progression, including response to gefitinib and captures tumor heterogeneity, patient age, gender, initial clinical disease stage, and smoking status. Methods: 5-step in silico model development: Model building: biology of EGFR+LUAD was characterized by extracting biological features and their functional relationships from > 300 published papers and translating them into ordinary differential equations (ODEs). Mutational burden, EGFR-downstream-pathways, tumor growth and heterogeneity, gefitinib-PK/PD, treatment-induced resistance and clinical outcome were incorporated into a knowledge-based model containing 27-97 variables, 108-258 parameters and 13-83 ODEs reflecting intra-tumor clonal heterogeneity in a mechanistic manner. Calibration: published spheroid, xenograft and clinical data were used for stepwise calibration to find the correct parameter values. Relevant virtual populations (VPOPs) matching real patients baseline characteristics were generated for model benchmarking and validation. Benchmarking against a published data-based model: coverage of experimental interquartile range (IQR) with simulated IQR (precision) assesses model fit with experimental data, coverage of simulated IQR with experimental IQR (overlap) assesses model fit with experimental variability. Validation: a VPOP with comparable baseline characteristics was simulated and results compared against a published patient dataset that wasn’t used in any step during calibration. Matching of prediction over clinical data was done using both coverage and bootstrapped log-rank metrics. Results: Our model provides comparable outputs to the data-based model without having access to the exact original data (our model: precision of 62%, overlap of 91% VS data-based model: precision of 72% and an overlap of 86%). Besides, as a validation criterion, our model also successfully reproduces the time to progression observed in an independent clinical trial (coverage > 99%, negative log-rank tests > 98%). Conclusions: We simulated tumor growth and treatment response in advanced EGFR+ LUAD patients and successfully validated results both against an existing data-based model and a published clinical data. Our model highlights the potential of in silico modeling to better understand complex diseases progression and support efficient development of innovative therapies.

High-throughput tuning of ovarian cancer spheroids for on-chip invasion assays

We developed an invasion assay by using microfabricated culture devices. First, ovarian tumor spheroids were generated with a culture patch device consisting of an agarose membrane formed with a honeycomb microframe – the patch – and gelatin nanofiber backbone. By changing the dimensions of the honeycomb compartments we were able to control the number of cells and size of the spheroids. When the spheroids were placed on a patch coated with a thin membrane of fibrillary type I collagen, spheroid disruption was observed due to substrate induced cell migration. This process is straightforward and should be applicable to other cancer types, as well as assays under microfluidic conditions, thereby holding the potential for use in tumor modeling and anti-cancer drug development. • In vitro models are necessary for cancer research. • Generation of spheroids is key for predicting tumor growth and microenvironmental dependencies. • A robust strategy to control the size and number of cells in spheroids playing with dimensions of microfabricated supports. • We can induce spheroid disruption and invasion using support mimicking the extracellular matrix. • Supports integration in microfluidic chip to control independently metastasis size, microenvironment, drug circulation.

Investigation on tumor cells growth by Taguchi method

It is essential to understand and find the features that influence tumor growth and cancer stem cells, the crucial incurable factor in cancer treatment, and cancer's reversibility. Several critical cellular features include symmetric and asymmetric stem cells, migration, cellular quiescence, senescence, apoptosis, tumor location, and cell phenotype. Some other features also exist that can influence cancer disease. The most valuable features are extracted to provide an efficient and realistic agent-based model for predicting tumor growth. According to the Taguchi method, we use orthogonal tables and signal-to-noise ratio to determine the most effective characteristics and help us propose and evaluate our agent-based model better and more accurately. This practical and straightforward method allows us to consider different states that change cancer stem cells and tumor growth. We find that migration, location of the tumor, and cell senescence are the three most valuable features.

Machine learning mechanism for segmentation, progressive assessment and prediction of brain tumor growth

A brain tumor is one type of illness caused by clots in the brain. A Magnetic Resonance Imaging (MRI) scan can be used to see a brain tumor in detail. Because of the similarity in color, it is difficult to distinguish brain tumor tissue from normal tissue. Brain tumors are common in all age groups, early and accurate identification of tumor type is critical for determining the most successful treatment regimens for each patient's situation. In today's world, computer interpretation of medical images plays a major role in medical diagnosis. Brain tumors can be accurately diagnosed with the use of advanced technologies. Early-stage diagnoses, as well as properly quantifying the amount and intensity of sickness, which were previously a barrier for medical science, are now eliminated by recent technologies. The automated technique for brain tumor identification was used in the study. The technique includes gray-scale conversion for reducing computation requirements. Filter operation was used to eliminate unwanted noises as much as possible to assist in improved segmentation. Next brain tumor segmentation, isolates tumor tissue from surrounding edema, fat, and cerebrospinal fluid. K-means clustering is used for this segmentation. 

Effect of WW Domain-Containing Oxidoreductase Gene Polymorphism on Clinicopathological Characteristics of Patients with EGFR Mutant Lung Adenocarcinoma in Taiwan.

Lung adenocarcinoma is the most common histological type of non-small cell lung cancer, which accounts for the majority of lung cancers. Previous studies have showed that dysregulation of WW domain-containing oxidoreductase (WWOX) participates in the generation of several cancer types, including lung cancer. However, whether these WWOX polymorphisms are related to the clinical risk of epidermal growth factor receptor (EGFR)-mutated lung adenocarcinoma is worthy of investigation. The present study examined the relationship between the WWOX single-nucleotide polymorphisms (SNPs; rs11545028, rs12918952, rs3764340, rs73569323, and rs383362) and the clinicopathological factors in lung adenocarcinoma patients with or without EGFR mutations. We found that there was no significant difference in the genotype distribution of WWOX polymorphism between EGFR wild-type and EGFR mutant in patients with lung adenocarcinoma. Our results demonstrated that the presence of at least one G genotype (CG and GG) allele on WWOX rs3764340 was associated with a significantly higher risk of nearby lymph node involvement in those patients harboring EGFR mutations (odds ratio (OR) = 3.881, p = 0.010) compared with the CC genotype. Furthermore, in the subgroup of lung adenocarcinoma patients with the EGFR-L858R mutation, both WWOX rs3764340 C/G (OR = 5.209, p = 0.023) and rs73569323 C/T polymorphisms (OR = 3.886, p = 0.039) exhibited significant associations with the size of primary tumors and the invasion of adjacent tissues. In conclusion, these data indicate that WWOX SNPs may help predict tumor growth and invasion in patients with EGFR mutant lung adenocarcinoma, especially those with the EGFR-L858R mutant in Taiwan.

19P Evaluation of different pharmacokinetic-pharmacodynamic models for tumor growth delay prediction after chemotherapy

In preclinical studies to rank anticancer agents for their efficacy the delay in tumor growth after impermanent drug exposure is often used. For a more detailed description of the anticancer drug action, mathematical pharmacokinetics/pharmacodynamics (PK/PD) models were developed. Such models allow predicting tumor growth time evolution and particularly the delay for various drug dosages. Each PK/PD model is based on the mathematical formulation of the biological processes underlying the antitumor drug action.

Endocan serum levels in patients with low- and high-grade meningiomas: does this biomarker have an indicative role?

BackgroundMeningiomas are one of the most common tumors of the brain and central nervous system. The key role of endocan in predicting tumor growth and prognosis has been shown for several types of cancers; however, this role in meningiomas has not been evaluated. In the current study, we investigated the relationship between endocan serum levels with low- and high-grade meningiomas.ResultsThe serum level of endocan in the group with meningiomas was 283.34 (242.09-358.70) pg/ml and in the control group was 250.29 (207.56-329.71) pg/ml respectively (P = 0.172). Afterwards, patients were divided into three different groups (grades I, II, and III) and compared to the control. The level of endocan in the group with grade I of meningioma showed no significant difference compared to control individuals (P = 0.86). When patients with grade II and grade III compared with the control group, endocan serum levels were statistically significant (P = 0.002, P < 0.001 respectively). Moreover, our findings showed that the different grades of meningiomas were statistically significant compared to each other (P < 0.001) regarding endocan serum levels, meaning that the higher the grade, the higher the endocan serum levels.ConclusionOur findings revealed that higher grades of meningioma had higher endocan serum levels, however, the role of endocan in pathogenesis or progression of this type of tumor requiring further exclusively assessment.

Biologically-Based Mathematical Modeling of Tumor Vasculature and Angiogenesis via Time-Resolved Imaging Data.

Simple SummaryThe recruitment of new vasculature via angiogenesis is a critical component of tumor development, which fundamentally influences tumor growth and response to treatment. The characterization of tumor-induced angiogenesis via mathematical models could enable approaches to forecast tumor response and improve patient care. In this review, we discuss how time-resolved imaging data integrated with mathematical modeling can be used to systematically investigate angiogenesis from the cell to tissue scale and ultimately forecast response to therapy.Tumor-associated vasculature is responsible for the delivery of nutrients, removal of waste, and allowing growth beyond 2–3 mm3. Additionally, the vascular network, which is changing in both space and time, fundamentally influences tumor response to both systemic and radiation therapy. Thus, a robust understanding of vascular dynamics is necessary to accurately predict tumor growth, as well as establish optimal treatment protocols to achieve optimal tumor control. Such a goal requires the intimate integration of both theory and experiment. Quantitative and time-resolved imaging methods have emerged as technologies able to visualize and characterize tumor vascular properties before and during therapy at the tissue and cell scale. Parallel to, but separate from those developments, mathematical modeling techniques have been developed to enable in silico investigations into theoretical tumor and vascular dynamics. In particular, recent efforts have sought to integrate both theory and experiment to enable data-driven mathematical modeling. Such mathematical models are calibrated by data obtained from individual tumor-vascular systems to predict future vascular growth, delivery of systemic agents, and response to radiotherapy. In this review, we discuss experimental techniques for visualizing and quantifying vascular dynamics including magnetic resonance imaging, microfluidic devices, and confocal microscopy. We then focus on the integration of these experimental measures with biologically based mathematical models to generate testable predictions.

Clinical and radiological findings for small renal masses under active surveillance

ObjectiveTo analyze the experience performing active surveillance (AS) of small renal masses (SRMs) in our center and to correlate the evolution of SRMs under AS with clinical and radiological findings. MethodsPatients on AS between January 2012 until May 2020 for SRMs in our center have been included. Growth rate (GR) per year was analyzed and correlated with radiographic features. Patients with growth kinetics higher than 5mm/year during follow up were offered active treatment. Results73 patients were included in AS: the mean age was 75.7 years, a mean initial tumour size of 21.2 mm, and a mean growth rate of 2.05 mm/year. Around 60 % had an ASA score of 3. The tumor size did not change over time in 43% of cases; in 4% we noticed a regression in size and in 52% of cases growth during follow-up (38% 1-5mm/year and 14% more than 5 mm/year).Delayed active treatment was indicated in 16 (21%) of cases. Treatment applied was as following: 2 radiofrequency ablations, 6 radical and 8 partial nephrectomies.A weak correlation was found between initial size and growth rate (r = 0.38, P = 0.02). No significant association was detected regarding any of the analyzed radiological findings and GR.With a mean follow up time of 33 months none of the patients presented metastatic progression. ConclusionActive surveillance is a feasible option for management of SRMs in selected patients without jeopardizing oncological safety. In our series, no clinical or radiological characteristics for predicting tumour growth were found.

Modeling of Glioma Growth With Mass Effect by Longitudinal Magnetic Resonance Imaging.

It is well-known that expanding glioblastomas typically induce significant deformations of the surrounding parenchyma (i.e., the so-called "mass effect"). In this study, we evaluate the performance of three mathematical models of tumor growth: 1) a reaction-diffusion-advection model which accounts for mass effect (RDAM), 2) a reaction-diffusion model with mass effect that is consistent only in the case of small deformations (RDM), and 3) a reaction-diffusion model that does not include the mass effect (RD). The models were calibrated with magnetic resonance imaging (MRI) data obtained during tumor development in a murine model of glioma (n = 9). We obtained T2-weighted and contrast-enhanced T1-weighted MRI at 6 time points over 10 days to determine the spatiotemporal variation in the mass effect and the volume fraction of tumor cells, respectively. We calibrated the three models using data 1) at the first four, 2) only at the first and fourth, and 3) only at the third and fourth time points. Each of these calibrations were run forward in time to predict the volume fraction of tumor cells at the conclusion of the experiment. The diffusion coefficient for the RDAM model (median of 10.65 × 10 -3 mm 2· d -1) is significantly less than those for the RD and RDM models (17.46 × 10 -3 mm 2· d -1 and 19.38 × 10 -3 mm 2· d -1, respectively). The error in the tumor volume fraction for the RD, RDM, and RDAM models have medians of 40.2%, 32.1%, and 44.7%, respectively, for the calibration using data from the first four time points. The RDM model most accurately predicts tumor growth, while the RDAM model presents the least variation in its estimates of the diffusion coefficient and proliferation rate. This study demonstrates that the mathematical models capture both tumor development and mass effect observed in experiments.

Predicting the effectiveness of chemotherapy using stochastic ODE models of tumor growth

Ordinary differential equation (ODE) models of cancer growth are often used to predict tumor growth and form the basis for more complex models used in personalized medicine. Unfortunately, ODE models provide predictions of the average behavior of the cell population neglecting the fact that cells are discrete objects subject to discrete events. This kind of stochasticity can dramatically change the time course of tumor growth, particularly when the cell population is small. Here, we investigate stochastic versions of seven common ODE models of cancer growth to determine the role of stochasticity in eradicating tumors via chemotherapy. We find that stochasticity leads to differences in predictions among the different models of both the level of chemotherapy needed to cure a tumor and the time it takes to achieve a cure. Our results highlight the need for more investigation of which model provides the best description of cancer growth.

Mechanistic insights into the heterogeneous response to anti‐VEGF treatment in tumors

AbstractVascular endothelial growth factor (VEGF) is a strong promoter of angiogenesis in tumors, and anti‐VEGF treatment, such as a humanized antibody to VEGF, is clinically used as a monotherapy or in combination with chemotherapy to treat cancer patients. However, this approach is not effective in all patients or cancer types. To better understand the heterogeneous responses to anti‐VEGF and the synergy between anti‐VEGF and other anticancer therapies, we constructed a computational model characterizing angiogenesis‐mediated growth of in vivo mouse tumor xenografts. The model captures VEGF‐mediated cross‐talk between tumor cells and endothelial cells and is able to predict the details of molecular‐ and cellular‐level dynamics. The model predictions of tumor growth in response to anti‐VEGF closely match the quantitative measurements from multiple preclinical mouse studies. We applied the model to investigate the effects of VEGF‐targeted treatment on tumor cells and endothelial cells. We identified that tumors with lower tumor cell growth rate and higher carrying capacity have a stronger response to anti‐VEGF treatment. The predictions indicate that the variation of tumor cell growth rate can be a main reason for the experimentally observed heterogeneous response to anti‐VEGF. In addition, our simulation results suggest a new synergy mechanism where anticancer therapy can enhance anti‐VEGF simply through reducing the tumor cell growth rate. Overall, this work generates novel insights into the heterogeneous response to anti‐VEGF treatment and the synergy of anti‐VEGF with other therapies, providing a tool that be further used to test and optimize anticancer therapy.

Lung Tumor Growth Modeling in Patients with NSCLC Undergoing Radiotherapy

This paper proposes two modeling approaches to predict lung tumor dynamics as an effect of radiotherapy. Real clinical information of non-small cell lung cancer (NSCLC) patients undergoing stereotactic body radiation therapy (SBRT) as the primary treatment method has been used for numerical simulations. The classical Gompertz model for tumor volume growth prediction was modified using a fractional parameter and combined with the linear-quadratic model to foresee the effect of SBRT on the targeted tumor. Another approach was implemented by following a pharmacokinetic-pharmacodynamic (PKPD) minimal compartmental model for single therapy with SBRT. Statistical analysis has been carried out to compare the two models. In terms of tumor growth prediction, obtained results indicated a decrease in the total tumor volume for both modeling approaches. A striking observation to emerge from the data comparison is the interesting perspective of fractional tools for further exploration in modeling tumor growth. Copyright (C) 2021 The Authors.

Expression of CEACAM1 and CD105 in Renal Cell Carcinoma and Its Correlation with Microvessel Density.

We investigated expression of carcinoembryonic antigen-related cell adhesion molecule 1 (CEACAM1) and endoglin (CD105) in renal cell carcinoma (RCC), and its potential role in predicting tumor growth and progression. A total of 47 RCC specimens and 15 adjacent normal kidney tissues were obtained. Expression of CEACAM1 and CD105 was assessed by immunohistochemistry. Microvessel density (MVD) was counted under the microscope by labeling the endothelial cells with biomarker CD34. The positivity of CEACAM1 expression in RCC (42.6%) was significantly lower than that in the normal kidney (73.%, P = 0.038). In contrast, the positivity of CD105 expression was significantly higher in RCC (78.7%) compared to that in the normal kidney (46.7%, P = 0.017). The expression level of CD105 in 47 RCC patients was significantly associated with the clinical stages of RCC (P < 0.05) but not with gender, age, tumor size, or histologic grade. Average MVD in RCC (78.05 ± 16.57) was significantly higher than that in normal tissue (43.62 ± 12.37, P < 0.05), and was significantly higher in RCC patients with advanced histologic grades (P < 0.05) or clinical stages (P < 0.01). In addition, MVD was significantly correlated with CD105 but negatively correlated with CEACAM1. Our findings suggest that down-regulation of CEACAM1 may promote angiogenesis in RCC, and that up-regulation of CD105 may promote RCC progress. MVD may be an indicator of RCC malignancy.

Integrating Quantitative Assays with Biologically Based Mathematical Modeling for Predictive Oncology.

SummaryWe provide an overview on the use of biological assays to calibrate and initialize mechanism-based models of cancer phenomena. Although artificial intelligence methods currently dominate the landscape in computational oncology, mathematical models that seek to explicitly incorporate biological mechanisms into their formalism are of increasing interest. These models can guide experimental design and provide insights into the underlying mechanisms of cancer progression. Historically, these models have included a myriad of parameters that have been difficult to quantify in biologically relevant systems, limiting their practical insights. Recently, however, there has been much interest calibrating biologically based models with the quantitative measurements available from (for example) RNA sequencing, time-resolved microscopy, and in vivo imaging. In this contribution, we summarize how a variety of experimental methods quantify tumor characteristics from the molecular to tissue scales and describe how such data can be directly integrated with mechanism-based models to improve predictions of tumor growth and treatment response.

Physiologically-Based Pharmacokinetic/Pharmacodynamic Model of MBQ-167 to Predict Tumor Growth Inhibition in Mice.

MBQ-167 is a dual inhibitor of the Rho GTPases Rac and Cdc42 that has shown promising results as an anti-cancer therapeutic at the preclinical stage. This drug has been tested in vitro and in vivo in metastatic breast cancer mouse models. The aim of this study is to develop a physiologically based pharmacokinetic/pharmacodynamic (PBPK-PD) model of MBQ-167 to predict tumor growth inhibition following intraperitoneal (IP) administration in mice bearing Triple Negative and HER2+ mammary tumors. PBPK and Simeoni tumor growth inhibition (TGI) models were developed using the Simcyp V19 Animal Simulator. Our developed PBPK framework adequately describes the time course of MBQ-167 in each of the mouse tissues (e.g., lungs, heart, liver, kidneys, spleen, plasma) and tumor, since the predicted results were consistent with the experimental data. The developed PBPK-PD model successfully predicts tumor shrinkage in HER2+ and triple-negative breast tumors after the intraperitoneal administration of 1 and 10 mg/kg body weight (BW) dose level of MBQ-167 three times a week. The findings from this study suggest that MBQ-167 has a higher net effect and potency inhibiting Triple Negative mammary tumor growth compared to HER2+ and that liver metabolism is the major route of elimination of this drug.

Mathematical models in nursing research.

This paper discusses the use of advanced mathematical tools in nursing research, such as mathematical models used in medicine for description and prediction of experimental tumor growth. They are rarely used in nursing research, but fortunately in the last decade, their use is increased, mainly due to artificial intelligence and Big Data, with great benefits for further nursing development. Therefore, a strong interaction between nurses and mathematicians is needed to improve nursing research, and consequently, the nurses’ performance in daily work.Significance for public healthThe study described in this paper is significant for public health, because it discusses the importance of using mathematical models in nursing research. Mathematical models used in many scientific areas could impact nurses' daily work, guiding their decisions and helping them choose better strategies, resulting in an improvement in their performance.

GP-GAN: Brain tumor growth prediction using stacked 3D generative adversarial networks from longitudinal MR Images

Brain tumors are one of the major common causes of cancer-related death, worldwide. Growth prediction of these tumors, particularly gliomas which are the most dominant type, can be quite useful to improve treatment planning, quantify tumor aggressiveness, and estimate patients’ survival time towards precision medicine. Studying tumor growth prediction basically requires multiple time points of single or multimodal medical images of the same patient. Recent models are based on complex mathematical formulations that basically rely on a system of partial differential equations, e.g. reaction diffusion model, to capture the diffusion and proliferation of tumor cells in the surrounding tissue. However, these models usually have small number of parameters that are insufficient to capture different patterns and other characteristics of the tumors. In addition, such models consider tumor growth independently for each subject, not being able to get benefit from possible common growth patterns existed in the whole population under study. In this paper, we propose a novel data-driven method via stacked 3D generative adversarial networks (GANs), named GP-GAN, for growth prediction of glioma. Specifically, we use stacked conditional GANs with a novel objective function that includes both l1 and Dice losses. Moreover, we use segmented feature maps to guide the generator for better generated images. Our generator is designed based on a modified 3D U-Net architecture with skip connections to combine hierarchical features and thus have a better generated image. The proposed method is trained and tested on 18 subjects with 3 time points (9 subjects from collaborative hospital and 9 subjects from BRATS 2014 dataset). Results show that our proposed GP-GAN outperforms state-of-the-art methods for glioma growth prediction and attain average Jaccard index and Dice coefficient of 78.97% and 88.26%, respectively.

Epithelial/mesenchymal heterogeneity of high-grade serous ovarian carcinoma samples correlates with miRNA let-7 levels and predicts tumor growth and metastasis.

Patient-derived samples present an advantage over current cell line models of high-grade serous ovarian cancer (HGSOC) that are not always reliable and phenotypically faithful models of invivo HGSOC. To improve upon cell line models of HGSOC, we set out to characterize a panel of patient-derived cells and determine their epithelial and mesenchymal characteristics. We analyzed RNA and protein expression levels in patient-derived xenograft (PDX) models of HGSOC, and functionally characterized these models using flow cytometry, wound healing assays, invasion assays, and spheroid cultures. Besides invitro work, we also evaluated the growth characteristics of PDX invivo (orthotopic PDX). We found that all samples had hybrid characteristics, covering a spectrum from an epithelial-to-mesenchymal state. Samples with a stronger epithelial phenotype were more active in self-renewal assays and more tumorigenic in orthotopic xenograft models as compared to samples with a stronger mesenchymal phenotype, which were more migratory and invasive. Additionally, we observed an inverse association between microRNA let-7 (lethal-7) expression and stemness, consistent with the loss of let-7 being an important component of the cancer stem cell phenotype. We observed that lower let-7 levels were associated with the epithelial state and a lower epithelial mesenchymal transition (EMT) score, more efficient spheroid and tumor formation, and increased sensitivity to platinum-based chemotherapy. Surprisingly, in these HGSOC cells, stemness could be dissociated from invasiveness: Cells with lower let-7 levels were more tumorigenic, but less migratory, and with a lower EMT score, than those with higher let-7 levels. We conclude that let-7 expression and epithelial/mesenchymal state are valuable predictors of HGSOC proliferation, invitro self-renewal, and tumor burden invivo.

Abstract 1682: Epithelial-mesenchymal heterogeneity of high-grade serous ovarian carcinoma samples correlates with let-7 levels and predicts tumor growth and metastasis

Abstract Objective: Patient-derived samples present an advantage over current cell line models of high grade serous ovarian cancer (HGSOC), which are flawed in terms of being reliable and phenotypically faithful models of in vivo HGSOC. To improve upon cell line models of HGSOC, we characterized a panel of patient-derived cells to determine their epithelial and mesenchymal characteristics, invasiveness, proliferation, stemness, and in vivo growth. Experimental procedures: Patient-derived xenograft (PDX) models of HGSOC were analyzed in vitro for phenotypic (RNA, let-7 miRNA, and protein expression, flow cytometry) and functional aspects including growth in spheroids, wound healing assays, invasion assays and in vivo (orthotopic PDX) growth characteristics. Results: Samples fell along the spectrum from epithelial to mesenchymal, and all had hybrid characteristics. Those toward the more epithelial end of the spectrum were most active in self-renewal assays, and grew most robustly in orthotopic xenograft models. Chemoresistance correlated both with the mesenchymal state and with BRCA2 wild type status. Loss of microRNA let-7 (lethal-7) is an important component of the cancer stem cell phenotype, and we observed an inverse association between let-7 expression and the epithelial state. We observed lower levels of let-7, more efficient spheroid and tumor formation, and increased sensitivity to platinum-based chemotherapy in cells with the most epithelial phenotype. Conclusions: Surprisingly, in these HGSOC cells, stemness could be dissociated from invasiveness: epithelial cells (those with least let-7 expression) were less migratory, but more tumorigenic, than the mesenchymal cells with higher let-7 expression. We conclude that epithelial/mesenchymal state and let-7 expression are valuable predictors of HGSOC proliferation, in vitro self-renewal, and tumor burden in vivo. Citation Format: Juli Unternaehrer, Evgeny Chirshev, Nozomi Hojo, Antonella Bertucci, Linda Sanderman, Anthony Nguyen, Hanmin Wang, Tise Suzuki, Emmanuel Brito, Shannalee Martine, Christine Castañón, Saied Mirshahidi, Marcelo Vazquez, Kerby C. Oberg, Yevgeniya J. Ioffe. Epithelial-mesenchymal heterogeneity of high-grade serous ovarian carcinoma samples correlates with let-7 levels and predicts tumor growth and metastasis [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 1682.