INTRODUCTION: Cutibacterium acnes (C. acnes) is the most common organism associated with prosthetic joint infections (PJIs) in shoulder arthroplasty (SA). It usually manifests as a "stealth infection" that mimics aseptic loosening. Currently, the most common method for diagnosing C. acnes infections is culture, which takes approximately 14 days and has a 15-20% false-positive rate. This study aims to assess the efficacy of Raman spectroscopy in detecting C. acnes infection based on biofilm chemistry. METHODS: An analysis was conducted on biofilms grown on Ti6Al4V alloy discs, cultured from eight isolates: three strains each of C. acnes and C. modestum, and one strain each of S. epidermidis and S. aureus. Bacterial isolates were cultured under anaerobic conditions in Brain-Heart-Infusion broth (BHI) followed by 5-day incubation and fixation in 10% formalin. Raman spectra were cleaned of fluorescence in LabSpec, then post-processed in MATLAB using Savitzky-Golay smoothing and min-max normalization. Biofilm thickness was measured using laser-scanning confocal microscopy (VK-X3000, Keyence). Raman scans were also performed on surgically retrieved humeral stems. RESULTS: Biofilms exhibited a characteristic Raman signature distinct from bone, blood, and joint fluid, consisting of phenylalanine (~1004 cm⁻¹), amide III (~1250 cm⁻¹), CH₂/CH₃ (~1450 cm⁻¹), and amide I (~1650 cm⁻¹) bands. Signal quality correlated with biofilm thickness: the higher the biofilm, the higher the signal-to-noise ratio. The exploratory in vitro MLP classifier reached 86.7% overall accuracy. On clinical retrieval spectra, the triage stage reached 92% overall accuracy. In a clinical cohort of culture-confirmed patients, 11 of 12 patients agreed with culture on the presence of infection. CONCLUSIONS: By overcoming the 14-day turnaround and 15-20% false-positive rate of standard culture, this two-stage Raman/machine-learning pipeline offers a non-destructive, rapid diagnostic tool that can improve the detection of stealth C. acnes periprosthetic joint infections and support faster, more precise clinical decision-making.
INTRODUZIONE: Il Cutibacterium acnes (C. acnes) è l'organismo più comunemente associato alle infezioni periprotesiche (PJI) nell'artroplastica di spalla (SA). Si manifesta solitamente come un'"infezione silente" che imita l'allentamento asettico. Attualmente, il metodo più comune per diagnosticare le infezioni da C. acnes è la coltura, che richiede circa 14 giorni e presenta un tasso di falsi positivi del 15-20%. Questo studio si propone di valutare l'efficacia della spettroscopia Raman nel rilevare l'infezione da C. acnes basandosi sulla chimica del biofilm. METODI: È stata condotta un'analisi su biofilm cresciuti su dischi in lega di Ti6Al4V, coltivati a partire da otto isolati: tre ceppi ciascuno di C. acnes e C. modestum, e un ceppo ciascuno di S. epidermidis e S. aureus. Gli isolati batterici sono stati coltivati in condizioni anaerobiche in brodo Brain-Heart-Infusion (BHI), seguiti da un'incubazione di 5 giorni e fissazione in formalina al 10%. Gli spettri Raman sono stati depurati dalla fluorescenza in LabSpec, quindi post-elaborati in MATLAB utilizzando lo smoothing di Savitzky-Golay e la normalizzazione min-max. Lo spessore del biofilm è stato misurato mediante microscopia confocale a scansione laser (VK-X3000, Keyence). Le scansioni Raman sono state eseguite anche su steli omerali espiantati chirurgicamente. RISULTATI: I biofilm hanno mostrato una firma Raman caratteristica, distinta da quella di osso, sangue e liquido articolare, costituita dalle bande della fenilalanina (~1004 cm⁻¹), dell'ammide III (~1250 cm⁻¹), dei gruppi CH₂/CH₃ (~1450 cm⁻¹) e dell'ammide I (~1650 cm⁻¹). La qualità del segnale era correlata allo spessore del biofilm: biofilm con spessori maggiori mostravano un rapporto segnale-rumore migliore. Il classificatore MLP esplorativo in vitro ha raggiunto un'accuratezza complessiva del 86,7%. Per gli spettri di espianti clinici, la fase di triage ha raggiunto un'accuratezza complessiva del 92%. In una coorte clinica di pazienti con conferma colturale, 11 su 12 sono risultati concordanti con la coltura in termini di presenza di infezione. CONCLUSIONI: Superando il tempo di attesa di 14 giorni e il tasso di falsi positivi del 15-20% della coltura standard, questa pipeline diagnostica Raman/machine-learning a due fasi offre uno strumento diagnostico rapido e non distruttivo, in grado di migliorare il rilevamento delle infezioni periprotesiche silenti da C. acnes e di supportare decisioni cliniche più rapide e precise.
Raman spectroscopy for the rapid detection of C. acnes biofilms associated with shoulder arthroplasty
MASSI, MATILDE
2026/2027
Abstract
INTRODUCTION: Cutibacterium acnes (C. acnes) is the most common organism associated with prosthetic joint infections (PJIs) in shoulder arthroplasty (SA). It usually manifests as a "stealth infection" that mimics aseptic loosening. Currently, the most common method for diagnosing C. acnes infections is culture, which takes approximately 14 days and has a 15-20% false-positive rate. This study aims to assess the efficacy of Raman spectroscopy in detecting C. acnes infection based on biofilm chemistry. METHODS: An analysis was conducted on biofilms grown on Ti6Al4V alloy discs, cultured from eight isolates: three strains each of C. acnes and C. modestum, and one strain each of S. epidermidis and S. aureus. Bacterial isolates were cultured under anaerobic conditions in Brain-Heart-Infusion broth (BHI) followed by 5-day incubation and fixation in 10% formalin. Raman spectra were cleaned of fluorescence in LabSpec, then post-processed in MATLAB using Savitzky-Golay smoothing and min-max normalization. Biofilm thickness was measured using laser-scanning confocal microscopy (VK-X3000, Keyence). Raman scans were also performed on surgically retrieved humeral stems. RESULTS: Biofilms exhibited a characteristic Raman signature distinct from bone, blood, and joint fluid, consisting of phenylalanine (~1004 cm⁻¹), amide III (~1250 cm⁻¹), CH₂/CH₃ (~1450 cm⁻¹), and amide I (~1650 cm⁻¹) bands. Signal quality correlated with biofilm thickness: the higher the biofilm, the higher the signal-to-noise ratio. The exploratory in vitro MLP classifier reached 86.7% overall accuracy. On clinical retrieval spectra, the triage stage reached 92% overall accuracy. In a clinical cohort of culture-confirmed patients, 11 of 12 patients agreed with culture on the presence of infection. CONCLUSIONS: By overcoming the 14-day turnaround and 15-20% false-positive rate of standard culture, this two-stage Raman/machine-learning pipeline offers a non-destructive, rapid diagnostic tool that can improve the detection of stealth C. acnes periprosthetic joint infections and support faster, more precise clinical decision-making.| File | Dimensione | Formato | |
|---|---|---|---|
|
2026_10_Massi_Executive Summary.pdf
accessibile in internet per tutti a partire dal 10/09/2029
Descrizione: executive summary
Dimensione
2.96 MB
Formato
Adobe PDF
|
2.96 MB | Adobe PDF | Visualizza/Apri |
|
2026_10_Massi_tesi.pdf
accessibile in internet per tutti a partire dal 10/09/2029
Descrizione: testo completo tesi
Dimensione
59.08 MB
Formato
Adobe PDF
|
59.08 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/264318