A pathologist-AI collaboration framework for enhancing diagnostic accuracies and efficiencies.

In pathology, the deployment of artificial intelligence (AI) in clinical settings is constrained by limitations in data collection and in model transparency and interpretability. Here we describe a digital pathology framework, nuclei.io, that incorporates active learning and human-in-the-loop real-time feedback for the rapid creation of diverse datasets and models. We validate the effectiveness of the framework via two crossover user studies that leveraged collaboration between the AI and the pathologist, including the identification of plasma cells in endometrial biopsies and the detection of colorectal cancer metastasis in lymph nodes. In both studies, nuclei.io yielded considerable diagnostic performance improvements. Collaboration between clinicians and AI will aid digital pathology by enhancing accuracies and efficiencies.

B-lymphoblastic leukemia with transient spontaneous remission in the setting of severe group A streptococcus infection.

Spontaneous remission of B-lymphoblastic leukemia (B-ALL) in the setting of viral and bacterial infections has been reported. Here, we present a case of B-ALL that showed a complete remission in the setting of group A streptococcal bacteremia. The patient was an 11-year-old boy who presented with a sore throat, right ear pain, and rhinorrhea. Prior to the diagnosis of B-ALL, he was diagnosed with streptococcal pharyngitis and received a single dose of dexamethasone and azithromycin. One day later, he was found to be pancytopenic and an immunophenotypically abnormal B-lymphoblastic population was detected comprising 0.6% and 16.8% of the peripheral blood and bone marrow cells, respectively. Though a diagnosis of B-ALL was highly suspected, blast percentage was <20% and the bone marrow showed relatively unremarkable trilineage hematopoiesis. On close monitoring, the suspected neoplastic population became undetectable by day 17 and the patient's complete blood count (CBC) completely normalized by day 46. On day 82, a peripheral blood smear demonstrated circulating blasts. Flow cytometry of a bone marrow aspirate revealed B-lymphoblastic leukemia accounting for 94% nucleated cells, consistent with the diagnosis of B-lymphoblastic leukemia. This case is of interest as less than 20 examples of spontaneous remission of B-ALL have been reported in the literature. As the case reported here relapsed and previously reported spontaneously remitting cases have uniformly relapsed, cases of B-ALL with spontaneous remission should be followed very closely for recurrence.

Long-standing donor diabetes and pathologic findings are associated with shorter allograft survival in recipients of kidney transplants from diabetic donors.

Approximately 6% of deceased kidney donors (DKDs) are diabetic; their kidneys may be associated with worse allograft survival, but published studies suggest that recipient diabetes status has a greater impact on mortality and survival. Since biopsy findings are the most common reason for organ discard, we sought to understand histologic and clinical factors that influence graft survival in patients who receive a kidney from a diabetic DKD. We retrospectively reviewed our institutional experience from 2005 to 2019, and re-evaluated pre-implantation and earliest post-transplant biopsies. Histologic findings were compared against a control cohort of non-diabetic DKD. Of 829 adult DKD transplants, 37 (4.5%) came from diabetic donors. There was no significant difference in diabetic vs. non-diabetic DKD graft survival for all-comers; however, when stratified by duration of donor diabetes, donor diabetes ≥6 years was associated with graft failure. In 25 patients with post-transplant biopsies available, diabetic DKD allografts had significantly greater non-glomerular chronic injury than non-diabetic DKD allografts. Moderate arteriolar hyalinosis (in 24%), moderate tubular atrophy and interstitial fibrosis (IFTA, in 36%), and diabetic glomerulopathy (in 24%) on early post-transplant biopsy were associated with allograft failure. Pre-implantation frozen section discrepancies were more common in long-standing donor diabetes, and arteriolar hyalinosis and IFTA scores on frozen accurately prognosticated graft loss. There was no morphologic improvement in lesions of diabetic nephropathy on short-term follow-up. In conclusion, donor diabetes ≥6 years, and histologic findings on frozen section and early post-transplant biopsy are associated with diabetic DKD allograft loss.

PBP4 Mediates ß-Lactam Resistance by Altered Function.

Penicillin binding protein 4 (PBP4) can provide high-level ß-lactam resistance in Staphylococcus aureus A series of missense and promoter mutations associated with pbp4 were detected in strains that displayed high-level resistance. We show here that the missense mutations facilitate the ß-lactam resistance mediated by PBP4 and the promoter mutations lead to overexpression of pbp4 Our results also suggest a cooperative interplay among PBPs for ß-lactam resistance.

PBP 4 Mediates High-Level Resistance to New-Generation Cephalosporins in Staphylococcus aureus.

Staphylococcus aureus is an important cause of both hospital- and community-associated methicillin-resistant S. aureus (MRSA) infections worldwide. ß-Lactam antibiotics are the drugs of choice to treat S. aureus infections, but resistance to these and other antibiotics make treatment problematic. High-level ß-lactam resistance of S. aureus has always been attributed to the horizontally acquired penicillin binding protein 2a (PBP 2a) encoded by the mecA gene. Here, we show that S. aureus can also express high-level resistance to ß-lactams, including new-generation broad-spectrum cephalosporins that are active against methicillin-resistant strains, through an uncanonical core genome-encoded penicillin binding protein, PBP 4, a nonessential enzyme previously considered not to be important for staphylococcal ß-lactam resistance. Our results show that PBP 4 can mediate high-level resistance to ß-lactams.