Knösel, Thomas

Abstract Spatial transcriptomics links gene expression to tissue architecture, providing a mechanistic view of cellular organization. Yet existing datasets cover few donors and miss the complexity of human disease. Experimental costs remain prohibitive, and large-scale profiling is impractically slow for population-level studies. Accurate computational methods are urgently needed. Predicting gene expression from standard histology, however, remains an open problem, as current approaches transfer poorly to unseen cohorts and diseases. Here, we present Phoenix, a (latent) flow matching generative model that infers pan-cancer spatially resolved single-cell gene expression with high accuracy. Phoenix analyzes treatment response in silico: Applied to 763 head and neck cancer patients, it identified three new spatial biomarkers that we validated across two cancers (breast cancer, n = 84; ovarian cancer, n = 157) and treatment regimens (platinum, trastuzumab). Phoenix generalizes beyond carcinomas: In a large sarcoma cohort (802 tissue microarray cores), it accurately predicted cell-type-specific signatures in held-out samples and captured chemotherapy-induced immune remodeling. Phoenix also extends across species: In a mouse model, it accurately predicted the expression of pancreatic cancer lineage markers and the mutant mKras^G12D allele in silico. In total, we evaluated Phoenix on over 10,000 patients. Our results establish virtual spatial transcriptomics as a scalable framework for studying tissue organization, therapeutic response, and disease mechanisms.

Background: Infection with SARS-CoV-2 has initially been known as a respiratory disease but in the course of the pandemic the understanding has emerged that severity is owing to fatal inflammatory responses apart from lung injury. In this context, endocrine disorders such as thyroiditis as well as pituitary dysfunction in addition to nonthyroidal illness syndrome have been described. Furthermore, angiotensin-converting enzyme 2 (ACE2), the SARS-CoV-2 cell receptor, has been detected in most endocrine tissues, including the thyroid gland. Objective: To evaluate histopathologic changes and compare thyroidal ACE2 protein expression in thyroid tissue from patients who died from severe COVID-19 with thyroid tissue from patients without SARS-CoV-2 infection in a retrospective case series. Furthermore, to assess and compare alterations in thyroid function tests (TFTs) between patients with or without SARS-CoV-2 infection as well as association of TFTs with the severity of the disease in a prospective cohort study. Methods: Thyroid tissue of deceased COVID-19 patients ( n = 23) was analyzed for histopathology and ACE2 expression by immunohistochemical staining. A total of 153 patients with confirmed SARS-CoV-2 were evaluated regarding TFTs and divided into a severe (intubation, intensive care treatment) and an intermediate group. Results: Thyroidal ACE2 expression was detected in 87% of the deceased COVID-19 patients. Normal thyroid tissue from patients without SARS-CoV-2 infection showed no ACE2 protein expression. Half of the severely ill COVID-19 patients had low free triiodothyronine (fT3) levels. Combination of low fT3 and thyrotropin (TSH) was associated significantly with deadly disease. Conclusion: The high percentage of positive ACE2 immunostaining in deceased patients compared with normal thyroid tissue of patients without SARS-CoV-2 infection suggests involvement of the thyroid in COVID-19, although further research will have to show the pathogenic role of thyroidal ACE2 in COVID-19. Abnormal fT3 and a TSH of ≤0.5 mU/L were associated with a fatal outcome in our severely ill SARS-CoV-2 patient cohort. Therefore, assessment of TFTs is crucial in the treatment of severely ill COVID-19 patients. Trial Registration: COVID-19 Registry of the LMU University Hospital Munich (CORKUM), WHO trial ID DRKS00021225.

Background: RET (rearranged during transfection) variants are the most prevalent oncogenic events in medullary thyroid cancer (MTC). In advanced disease, multi-tyrosine kinase inhibitors (MKIs) cabozantinib and vandetanib are the approved standard treatment irrespective of RET status. The actual outcome of patients with RET-positive MTC treated with MKIs is ill described. Methods: We here retrospectively determined the RET oncogene variant status with a targeted DNA Custom Panel in a prospectively collected cohort of 48 patients with advanced MTC treated with vandetanib and/or cabozantinib at four German referral centers. Progression-free survival (PFS) and overall survival (OS) probabilities were estimated using the Kaplan-Meier method. Results: In total, 44/48 (92%) patients had germline or somatic RET variants. The M918T variant was found in 29/44 (66%) cases. In total, 2/32 (6%) patients with a somatic RET variant had further somatic variants, while in 1/32 (3%) patient with a germline RET variant, additional variants were found. Only 1/48 (2%) patient had a pathogenic HRAS variant, and no variants were found in 3 cases. In first-line treatment, the median OS was 53 (95% CI (95% confidence interval), 32–NR (not reached); n = 36), and the median PFS was 21 months (12–39; n = 33) in RET-positive MTC patients. In second-line treatment, the median OS was 18 (13–79; n = 22), and the median PFS was 3.5 months (2–14; n = 22) in RET-positive cases. Conclusions: RET variants were highly prevalent in patients with advanced MTC. The treatment results in RET-positive cases were similar to those reported in unselected cohorts.

Summary Struma ovarii is a teratoma of the ovaries predominantly composed of thyroid tissue. Hyperthyroidism associated with struma ovarii is rare, occurring in approximately 8% of cases. Due to the rarity of struma ovarii, available data are limited to case reports and small case series.We report on a 61-year-old female patient with known Hashimoto’s thyroiditis on levothyroxine replacement therapy for years with transition to clinical and biochemical hyperthyroidism despite antithyroid medication with carbimazole (10 mg/day), new diagnosis of urothelial carcinoma and an adnexal mass suspicious of ovarian cancer. The patient underwent resection of the adnexal mass and histopathology revealed a mature teratoma predominantly composed of thyroid tissue showing high levels of sodium iodide symporter protein expression. Following struma ovarii resection and disappearance of autonomous production of thyroid hormones, the patient developed hypothyroidism with severely decreased thyroid hormone levels fT4 and fT3 (fT4 0.4 ng/dL, reference interval 0.9–1.7 and fT3 < 1.0 pg/mL, reference interval 2.0–4.4). This has previously been masked by continued thyroid-stimulating hormone suppression due to long-term hyperthyroidism pre-surgery indicating secondary hypothyroidism, in addition to primary hypothyroidism based on the known co-existing chronic lymphocytic thyroiditis of the orthotopic thyroid gland. Levothyroxine administration was started immediately restoring euthyroidism.This case illustrates possible diagnostic pitfalls in a patient with two concurrent causes of abnormal thyroid function. Learning points: Struma ovarii is an ovarian tumor containing either entirely or predominantly thyroid tissue and accounts for approximately 5% of all ovarian teratomas. In rare cases, both benign and malignant struma ovarii can secrete thyroid hormones, causing clinical and biochemical features of hyperthyroidism. Biochemical features of patients with struma ovarii and hyperthyroidism are similar to those of patients with primary hyperthyroidism. In such cases, thyroid scintigraphy should reveal low or absent radioiodine uptake in the thyroid gland, but the presence of radioiodine uptake in the pelvis in a whole body radioiodine scintigraphy. We give advice on possible diagnostic pitfalls in a case with two simultaneous causes of abnormal thyroid function due to the co-existence of struma ovarii.