MultiCOVID: a multi modal deep learning approach for COVID-19 diagnosis

The rapid spread of the severe acute respiratory syndrome coronavirus 2 led to a global overextension of healthcare. Both Chest X-rays (CXR) and blood test have been demonstrated to have predictive value on Coronavirus Disease 2019 (COVID-19) diagnosis on different prevalence scenarios. With the objective of improving and accelerating the diagnosis of COVID-19, a multi modal prediction algorithm (MultiCOVID) based on CXR and blood test was developed, to discriminate between COVID-19, Heart Failure and Non-COVID Pneumonia and healthy (Control) patients. This retrospective single-center study includes CXR and blood test obtained between January 2017 and May 2020. Multi modal prediction models were generated using opensource DL algorithms. Performance of the MultiCOVID algorithm was compared with interpretations from five experienced thoracic radiologists on 300 random test images using the McNemar-Bowker test. A total of 8578 samples from 6123 patients (mean age 66 ± 18 years of standard deviation, 3523 men) were evaluated across datasets. For the entire test set, the overall accuracy of MultiCOVID was 84%, with a mean AUC of 0.92 (0.89-0.94). For 300 random test images, overall accuracy of MultiCOVID was significantly higher (69.6%) compared with individual radiologists (range, 43.7-58.7%) and the consensus of all five radiologists (59.3%, P < .001). Overall, we have developed a multimodal deep learning algorithm, MultiCOVID, that discriminates among COVID-19, heart failure, non-COVID pneumonia and healthy patients using both CXR and blood test with a significantly better performance than experienced thoracic radiologists.

Evaluation of antimicrobial regimens in a guinea-pig model of meningitis caused by Pseudomonas aeruginosa

To compare the efficacy of meropenem, ceftazidime, tobramycin and ceftazidime+tobramycin in a guinea-pig model of P. aeruginosa meningitis. After anesthesia, the atlanto-occipital membrane was punctured with a butterfly needle and 100 microl of a solution containing 10(6)CFU/ml of P. aeruginosa were injected directly into the cisterna magna. Four h later, therapy was initiated with saline or antibiotics given im for 48 h in doses that obtained CSF levels as in human meningitis: ceftazidime 200 mg/kg/8h, meropenem 200 mg/kg/8h, tobramycin 30 mg/kg/24h. Tobramycin was also given intracisternally. Animals were sacrificed at different time points. CSF and blood samples were collected and a meningeal swab was performed. Four hours after inoculation, bacterial concentration in CSF was 4 to 5log10CFU and mean WBC was 16,000/-l. All control animals died in 24h with a 12% increase in cerebral edema. All blood-cultures were negative. Ceftazidime, ceftazidime+tobramycin and meropenem reduced the CSF bacterial concentration at 8h by 2.5log10. At 48 h all CSF cultures were sterile but meningeal swab cultures remained positive in 30%. Our results suggest that meropenem may be at least as effective as ceftazidime and that the addition of tobramycin to ceftazidime may improve its efficacy.

Experimental study of teicoplanin, alone and in combination, in the therapy of cephalosporin-resistant pneumococcal meningitis

OBJECTIVES

The aim of the study was to determine the efficacy of teicoplanin, alone and in combination with ceftriaxone, in a rabbit model of cephalosporin-resistant pneumococcal meningitis, and to assess the effect of concomitant therapy with dexamethasone.

METHODS

In vitro killing curves of teicoplanin, with and without ceftriaxone, were performed. Groups of eight animals per treatment were inoculated with a cephalosporin-resistant pneumococcal strain (penicillin MIC, 4 mg/L; ceftriaxone MIC, 2 mg/L; teicoplanin MIC, 0.03 mg/L) and treated over a 26 h period. Teicoplanin was administered at a dose of 15 mg/kg, alone and in combination with ceftriaxone at 100 mg/kg with or without dexamethasone at 0.25 mg/kg. CSF samples were collected at different time-points, and bacterial titres, white blood cell counts, lactate and protein concentrations and bacteriostatic/bactericidal titres were determined. Blood and CSF teicoplanin pharmacokinetic and pharmacodynamic parameters were determined.

RESULTS

Teicoplanin alone promoted a decrease in bacterial counts at 6 h of -2.66 log cfu/mL and was bactericidal at 24 h, without therapeutic failures. Similar good results were obtained when dexamethasone was used simultaneously, in spite of the penetration of teicoplanin into the CSF being significantly reduced, from 2.31% to 0.71%. Teicoplanin and ceftriaxone combinations were synergic in vitro, but not in the meningitis model.

CONCLUSIONS

Teicoplanin alone was very effective in this model of cephalosporin-resistant pneumococcal meningitis. The use of concomitant dexamethasone resulted in lower CSF teicoplanin levels, but not in therapeutic failures. The combination of teicoplanin plus ceftriaxone and dexamethasone might be a good alternative for the empirical therapy of pneumococcal meningitis. Additional data should confirm our experiments, in advance of clinical trials to assess efficacy in humans.