B. Beomonte Zobel

Experimental assessment of CT-based thermometry during laser ablation of porcine pancreas.

Laser interstitial thermotherapy (LITT) is employed to destroy tumors in organs, and its outcome strongly depends on the temperature distribution inside the treated tissue. The recent introduction of computed tomography (CT) scan thermometry, based on the CT number dependence of the tissue with temperature, overcomes the invasiveness of other techniques used to monitor temperature during LITT. The averaged CT number (ROI = 0.02 cm(2)) of an ex vivo swine pancreas is monitored during LITT (Nd:YAG laser power of 3 W, treatment time: 120 s) at different distances from the applicator (from 4 to 30 mm). The averaged CT number shows a clear decrease during treatment: it is highest at 4 mm from the applicator (mean variation in the whole treatment of -0.256 HU s(-1)) and negligible at 30 mm, since the highest temperature increase is present close to the applicator (i.e., 45 °C at 4 mm and 25 °C at 6 mm). To obtain the relationship between CT numbers and pancreas temperature, the reference temperature was measured by 12 fiber Bragg grating sensors. The CT number decreases as a function of temperature, showing a nonlinear trend with a mean thermal sensitivity of -0.50 HU °C(-1). Results here reported are the first assessment of pancreatic CT number dependence on temperature, at the best of our knowledge. Findings can be useful to further investigate CT scan thermometry during LITT on the pancreas.

Temperature monitoring during microwave ablation in ex vivo porcine livers

OBJECTIVE The aim of the present study was to assess the temperature map and its reproducibility while applying two different MWA systems (915 MHz vs 2.45 GHz) in ex vivo porcine livers. MATERIALS AND METHODS Fifteen fresh pig livers were treated using the two antennae at three different settings: treatment time of 10 min and power of 45 W for both systems; 4 min and 100 W for the 2.45 GHz system. Trends of temperature were recorded during all procedures by means of fiber optic-based probes located at five fixed distances from the antenna, ranging between 10 mm and 30 mm. Each trial was repeated twice to assess the reproducibility of temperature distribution. RESULTS Temperature as function of distance from the antenna can be modeled by a decreasing exponential trend. At the same settings, temperature obtained with the 2.45 GHz system was higher than that obtained with the 915 MHz thus resulting into a wider area of ablation (diameter 17 mm vs 15 mm). Both systems showed good reproducibility in terms of temperature distribution (root mean squared difference for both systems ranged between 2.8 °C and 3.4 °C). CONCLUSIONS When both MWA systems are applied, a decreasing exponential model can predict the temperature map. The 2.45 GHz antenna causes higher temperatures as compared to the 915 MHz thus, resulting into larger areas of ablation. Both systems showed good reproducibility although better results were achieved with the 2.45 GHz antenna.

Magnetic resonance-based thermometry during laser ablation on ex-vivo swine pancreas and liver

Laser Ablation (LA) is a minimally-invasive procedure for tumor treatment. LA outcomes depend on the heat distribution inside tissues and require accurate temperature measurement during the procedure. Magnetic resonance imaging (MRI) allows a non-invasive and three-dimensional thermometry of the organ undergoing LA. In this study, the temperature distribution within two swine pancreases and three swine livers undergoing LA (Nd:YAG, power: 2 W, treatment time: 4 min) was monitored by a 1.5-T MR scanner, utilizing two T1-weighted sequences (IRTF and SRTF). The signal intensity in four regions of interest, placed at different distances from the laser applicator, was related to temperature variations monitored in the same regions by twelve fiber Bragg grating sensors. The relationship between the signal intensity and temperature increase was calculated to obtain the calibration curve and to evaluate accuracy, sensibility and precision of each sequence. This is the first study of MR-based thermometry during LA on pancreas. More specifically, the IRTF sequence provides the highest temperature sensitivity in both liver (1.8 ± 0.2 °C(-1)) and pancreas (1.8 ± 0.5 °C(-1)) and the lowest precision and accuracy. SRTF sequence on pancreas presents the highest accuracy and precision (MODSFRT = -0.1 °C and LOASFRT = [-2.3; 2.1] °C).

E-learning in radiology: Italian multicentre experience = E-learning in radiologia: esperienza multicentrica italiana

PurposeThe aim of our work was to design, implement and evaluate an e-learning programme in favour of trainee radiologists enrolled at the many specialty schools located throughout Italy, in the spirit of “sharing culture”.Materials and methodsOnce a common educational programme and time slot had been identified and planned, the programme was delivered via Internet-based video conferencing once a week for 2-h lectures. Each lecture was followed by interaction between the teaching staff and trainees at the individual sites. The universities involved were Trieste, Udine, Verona, Milano Bicocca, Novara, Varese, Genova, Sassari, Rome “Campus”, Rome “Cattolica”, Chieti, Foggia, Catania, Modena and Firenze. The University of Rome “Cattolica” participated in the project with two locations: Rome “Policlinico Gemelli” and Rome “Cattolica Campus of Campobasso”.ResultsEighteen lectures were conducted, for a total of 36 h. “Transient” connection interruption occurred 13 times for a total of 33 min over 2,160 min of lessons. Video quality and, in particular, details of radiological images shown in slides or moving pictures, were rated as very good by 71% of trainees, good by 24% and satisfactory by 4.5%; no one gave a rating of unsatisfactory.ConclusionsBased on our experience, whereas e-learning in radiology has become established and compulsory, there is the need for legislation that on the one hand protects online teaching activity and on the other allows study and continuing medical education (CME) credits to be recognised.RiassuntoObiettivoScopo del nostro lavoro è stato progettare, realizzare e valutare un percorso didattico di e-learning per la formazione degli specializzandi in radiodiagnostica, coinvolgendo più scuole di specializzazione distribuite su tutto il territorio nazionale, nell’ottica della condivisione culturale.Materiali e metodiPrevia identificazione e pianificazione di un programma didattico temporale condiviso, la modalità tecnica di e-learning in teleconferenza ha previsto un collegamento via internet il mercoledÌ pomeriggio e due ore di lezione, alla lezione è seguita una interazione didattica delle singole sedi con il discente. Il ciclo didattico è iniziato il 21.10.2009 per terminare il 14.07.2010. Le Università che hanno aderito sono state: Trieste, Udine, Verona, Milano Bicocca, Novara, Varese, Genova, Sassari, Roma Campus, Roma Cattolica, Chieti, Foggia, Catania, Modena, Firenze. L’Università Cattolica di Roma ha partecipato al progetto con due sedi di collegamento: Roma Policlinico Gemelli e Roma Cattolica sede di Campobasso.RisultatiComplessivamente sono state effettuale 18 lezioni, per un totale di 36 ore di lezione effettive. Su un totale di 2160 minuti complessivi di collegamento le interruzioni momentanee della linea internet sono state 13 per un totale di 33 minuti. Relativamente al giudizio sulla qualità del video e, in particolare, sul dettaglio delle immagini radiologiche proposte in diapositiva e/o in filmati il 71% dei discenti ha espresso un giudizio “ottimo”, il 24% un giudizio “buono”, il 4,5% un giudizio “sufficiente” non vi è stato alcun giudizio “insufficiente”. Relativamente alla qualità audio il 72% dei discenti ha espresso un giudizio “ottimo”, il 25% un giudizio “buono”, il 3% un giudizio “sufficiente”, non vi è stato alcun giudizio “insufficiente”.ConclusioniDa quanto esposto e dall’esperienza vissuta se la via dell’e-learnig in radiologia è, a nostro parere, ormai segnata ed obbligatoria, si impone la necessità legislativa di normative che, da una parte, tutelino l’attività didattica in rete e, dall’altra, consentano il riconoscimento di crediti formativi e di educazione continua in medicina (ECM).

Monitoring of temperature increase and tissue vaporization during laser interstitial thermotherapy of ex vivo swine liver by computed tomography

Laser interstitial thermotherapy (LITT) is a minimally invasive technique used to thermally destroy tumour cells. Being based on hyperthermia, LITT outcome depends on the temperature distribution inside the tissue. Recently, CT scan thermometry, based on the dependence of the CT number (HU) on tissue temperature (T) has been introduced during LITT; it is an attractive approach to monitor T because it overcomes the concerns related to the invasiveness. We performed LITT on nine ex vivo swine livers at three different laser powers, (P=1.5 W, P=3 W, P=5 W) with a constant treatment time t=200 s; HU is averaged on two ellipsoidal regions of interest (ROI) of 0.2 cm2, placed at two distances from the applicator (d=3.6 mm and d=8.7 mm); a reference ROI was placed away from the applicator (d=30 mm). The aim of this study is twofold: 1) to evaluate the effect of the T increase in terms of HU variation in ex vivo swine livers undergoing LITT; and 2) to estimate the P value for tissue vaporization. To the best of our knowledge, this is the first study focused on the HU variation in swine livers undergoing LITT at different P. The reported findings could be useful to assess the effect of LITT on the liver in terms of both T changes and tissue vaporization, with the aim to obtain an effective therapy.

Infratentorial lesion volume correlates with sensory functional system in multiple sclerosis patients: a 3.0-Tesla MRI study

PurposeThis study sought to correlate lesion volume in infratentorial areas using 3.0-T proton-density (PD)-weighted images with disability scales and appropriate functional system scores in patients with multiple sclerosis (MS).Materials and methodsWe examined 20 consecutive patients (13 women and 7 men) with a median age of 47 years (range 26–70). Neurological examination included the Expanded Disability Status Scale and its functional systems, the Barthel Index (BI) and the Rivermead Mobility Index (RMI). MRI scans were performed on a system operating at 3.0 T using a quadrature birdcage head coil. Acquired images imported as Digital Imaging and Communication in Medicine (DICOM) files, and the region of interest (ROI) files were converted to Neuroimaging Informatics Technology Initiative (NIfTI) format and normalised to the Montreal Neurological Institute (MNI) standard template. An automated segmentation algorithm was used to distinguish between supratentorial and infratentorial areas. Normalisation to the magnetisation-prepared rapid acquisition with gradient echo (MPRAGE) T1-weighted sequence allowed lesion volume estimation in the different anatomical areas.ResultsA significant correlation was found between infratentorial lesion volume and the sensory functional system score (rho=0.76, p=0.002). No significant correlation was found between supratentorial lesion volume and Expanded Disability Status Scale (EDSS), RMI and BI scores.ConclusionsThe described method, by means of anatomical assignment of MS lesions, allows detection of significant correlation coefficients between clinical and MRI lesion burden in MS patients at the infratentorial level.RiassuntoObiettivoLo scopo del lavoro è di correlare il volume lesionale sotto-tentoriale in pazienti con sclerosi multipla (SM), utilizzando immagini di risonanza magnetica (RM) pesate in densità protonica a 3 Tesla, con scale di disabilità e sistemi funzionali.Materiali e metodiSono stati studiati 20 pazienti consecutivi (13 donne e 7 uomini) con una età mediana di 47 anni (intervallo 26–70 anni). L’esame neurologico prevedeva l’Expanded Disability Status Scale (EDSS) con i sistemi funzionali, il Barthel Index (BI) e il Rivermead Mobility Index (RMI). Gli esami di RM sono stati eseguiti utilizzando un magnete a 3 Tesla, con una bobina di quadratura per il cranio. Le immagini, importate come file tipo Digital Imaging and Communications in Medicine (DICOM) e file tipo Regions of Interest (ROI), sono state convertite in formato Neuroimaging Informatics Technology Initiative (NIfTI) e normalizzate rispetto al modello standard del Montreal Neurological Institute (MNI). Un algoritmo di segmentazione automatica è stato applicato per distinguere aree sopra- e sotto-tentoriali. La procedura di normalizzazione rispetto alla sequenza tipo magnetization-prepared rapid acquisition with gradient echo (MPRAGE) pesata in T1 ha permesso di stimare il volume di lesioni nelle differenti aree anatomiche.RisultatiÈ stata trovata una significativa correlazione tra il volume lesionale sotto-tentoriale e il punteggio ottenuto con il sistema funzionale sensoriale (rho=0,76, p=0,002). Non è stata osservata alcuna significativa correlazione tra til volume lesionale sopra-tentoriale ed EDSS, RMI e BI.ConclusioniIl metodo descritto, attraverso la attribuzione anatomica delle lesioni di SM, ha permesso di ottenere significativi coefficienti di correlazione tra la funzione sensoriale e il volume lesionale sotto-tentoriale misurato in RM in pazienti con SM.

DyCoH: an innovative tool to Dynamic Contrast Enhancement analysis

Contrast-Enhancement (CE) is an innovative approach, used in radiological framework, to evaluate the vascularization of the diseases. This non-invasive method determines the nature of a diseases, analysing the perfusions dynamic of contrast media in the tissues. In this paper we present an innovative tool named DyCoH (Dynamic Contrast Enhancement). This software, being specifically designed for this type of analysis, provides to medical doctor, in a very user-friendly framework, all the information needed to perform the CE analysis. DyCoH produces four inspectionable colour-maps that radiologists can use to identify the most relevant areas over which dynamically evaluates the contrast enhancement curve. However, the most interesting feature of DyCoH is its capability to manage, into a single framework, DICOM images produced by US, CT and MR of different vendors, allowing to support many types of clinical tests and to compare results provided by different diagnostic devices. Clinical tests have shown the effectiveness of the software and its capability to concretely support CE diagnoses.

Incidence of new fractures in women with osteoporosis-induced vertebral fractures detected on routine lateral chest radiographs

PurposeVertebral fractures represent one of the major complications of osteopororis. Diagnosis is followed by a pharmacological, interventional or surgical treatment. Up to day there are non practice guidelines for a screening evaluation of bone fractures in elderly and most of the fractures remain undiagnosed. We prospectively evaluated the prevalence of vertebral fractures on chest X-rays to determine the diagnostic and prognostic roles of chest X-ray in predicting new bone fractures 2 years after the initial radiogram.Materials and methodsBetween March 2004 and October 2005, 4,045 women underwent chest X-ray in our radiology department for any indication. We identified 166 women with the presence of at least one vertebral fracture. A questionnaire was administered to these women to collect information about diagnosis of osteoporosis, history of malignancy, systemic diseases, osteoporosis-inducing drugs and pharmacological, radiological or surgical treatment received.ResultsOut of the 166 women (age 73±10.5 years) with vertebral fractures, we interviewed 101 women; 13 had died and 52 were not found. Most of the patients were on menopause (97.1%, 98/101) with an average age of menopause of 48,2 years (±6 years). Among the patients on menopause, 15,8% (16/101) had undergone hysterectomy. All patients received a diagnosis of osteoporosis, which was reached with a chest X-ray report in 23.7% (24/101) of cases. A new skeletal fracture occurred in 20.5% (5/27) of patients receiving treatment against a frequency of 20.8% (16/74) in patients without treatment. No statistical difference was found between the groups (p = 0.374).ConclusionsInadequate treatment may explain the lack of a substantial difference in new fracture risk between treated and untreated patients. For these reason we discuss about the evaluation of an adeguate therapeutic approaches in prevention of osteoporosis-induced fractures.RiassuntoObiettvoLe fratture vertebrali rappresentano una delle maggiori complicanze dell’osteoporosi, la cui diagnosi è seguita dall’instaurazione di un trattamento medico-chirurgico. Attualmente non esistono linee guida che raccomandino uno screening nella popolazione di età avanzata e molte fratture rimangono non diagnosticate. L’obiettivo del nostro studio è quello di valutare prospetticamente la prevalenza delle fratture vertebrali visualizzate alla radiografia del torace e determinarne il ruolo diagnostico e prognostico calcolando il rischio di nuove fratture due anni dopo la prima diagnosi radiologica.Materiali e metodiDa marzo 2004 ad ottobre 2005, presso il nostro dipartimento di Diagnostica per Immagini, sono state eseguite 4045 radiografie del torace su donne che si sottoponevano a tale esame per svariate indicazioni. Dallo studio di tali referti sono state individuate 166 donne con fratture vertebrali secondarie ad osteoporosi. In base ai referti ottenuti, ci si è proposti di intervistare tali pazienti attraverso un questionario allo scopo di comprendere quante di loro fossero a conoscenza della malattia, di verificare l’esistenza di patologie concomitanti e di ricercare eventuali fattori che possano aver predisposto allo sviluppo delle lesioni vertebrali e l’eventuale terapia in atto.RisultatiDelle 166 pazienti individuate (età media 73 anni±10,5), con diagnosi di deformazioni o fratture vertebrali secondarie ad osteoporosi, sono state intervistate 101 donne; 13 sono decedute e 52 non sono state intervistate per difficoltà nel reperirle. La maggior parte di queste è risultata in menopausa (97,1%, 98/101) e l’età media della menopausa è stata di 48,2 anni (±6 anni). Inoltre, tra tutte le pazienti in menopausa, il 15,8% (16/101) ha subito intervento di isterectomia. Delle 101 pazienti al corrente della malattia, il 23,7% (24/101) ne è venuto a conoscenza solo in occasione della radiografia del torace da noi eseguita. Un nuovo evento fratturativo si è verificato nel 20,5% (5/27) delle pazienti in terapia contro una frequenza del 20,8% (16/74) delle pazienti che non avevano mai intrapreso alcun trattamento per l’osteoporosi. Pertanto non è stata riscontrata una differenza statisticamente significativa tra i due gruppi (p=0,374).ConclusioniUn trattamento inadeguato spiega l’assenza di differenza nel rischio di nuove fratture tra pazienti sottoposte o meno a terapia. Per tale ragione risulta necessario valutare le cause dell’assenza di un adeguato approccio terapeutico nella prevenzione delle fratture indotte dall’osteoporosi.

Differentiation of normal and neoplastic bone tissue in dynamic gadolinium-enhanced magnetic resonance imaging: validation of a semiautomated technique

PurposeThis study was undertaken to clinically validate the accuracy of a semiautomated software tool for analysing the enhancement curve in focal malignant bone lesions.Materials and methodsTwenty-three patients affected by cancer with malignant focal bone lesions underwent dynamic gadolinium-enhanced magnetic resonance (MR) imaging using the following protocol: T1-weighted turbo spin-echo sequences (time to repeat [TR] 600 ms, time to echo [TE] 8.6 ms, field of view [FOV] 40×40 cm) before and after intravenous injection of gadolinium-containing contrast agent. Image postprocessing was performed using the software DyCoH. Each region of interest (5×5 pixels), drawn to include the area of the lesion with the highest values of the area under the curve map, was analysed to obtain time-intensity curves and relative perfusion parameters: time to peak (TTP), peak intensity (PI), slope (60-s slope), intensity at 60 s after contrast agent injection (60-s I) and final intensity (FI).ResultsData were obtained by analysing 86 malignant lesions and 86 apparently normal bone regions. PI, 60-s slope, 60-s I and FI were significantly different between neoplastic and apparently normal (p<0.001) samples. Sensitivity, specificity and accuracy were, respectively, 94%, 93% and 94% at a PI threshold of 100 (signal-to-noise ratio), with positive and negative predictive values of 93% and 94%. At a threshold value of 0.85 for 60-s slope, sensitivity and specificity values were both 91%.ConclusionsThe semiautomated technique we report appears to be accurate for identifying neoplastic tissue and for mapping perfusion parameters, with the added value of a consistent measurement of perfusion parameters on colour-coded maps.RiassuntoObiettivoScopo del presente lavoro è stato valutare l’accuratezza di un software di analisi semiautomatica nell’analisi delle curve di potenziamento post-contrastografico delle lesioni ossee focali maligne.Materiali e metodiVentitre pazienti oncologici con lesioni ossee focali maligne sono stati sottoposti ad esame di risonanza magnetica dinamica utilizzando sequenze turbo spin echo T1-pesate (tempo di ripetizione [TR] 600 ms, tempo di eco [TE] 8,6 ms, campo di vista [FOV] 40×40 cm) prima e dopo somministrazione endovenosa di mezzo di contrasto paramagnetico. L’analisi delle immagini è stata poi effettuata utilizzando il software DyCoH. Ciascuna regione di interesse (ROI, 5×5 pixel), posizionata nella zona della lesione che presentava i più alti valori di area sotto la curva (AUC) sulla mappa, è stata analizzata per ottenere le curve intensità/tempo ed i relativi parametri perfusionali: tempo di picco (TTP), intensità di picco (PI), pendenza (60”Slope) ed intensità (60”I) al primo minuto dopo la somministrazione di mezzo di contrasto, intensità finale (FI).RisultatiI risultati sono stati ottenuti dall’analisi di 86 lesioni maligne ed 86 regioni di tessuto osseo apparentemente normale. I parametri PI, 60”Slope, 60”I e FI hanno mostrato valori con differenza statisticamente significativa tra lesioni neoplastiche e tessuto osseo apparentemente normale (p<0,001). La sensibilità, la specificità e l’accuratezza calcolate si sono mostrate rispettivamente del 94%, 93% e 94% ad un valore soglia di PI di 100 (signal to noise ratio, SNR), con valori predittivi positivi e negativi rispettivamente del 93% e 94%. Ad un valore soglia di 0,85 per il parametro 60”Slope, la sensibilità, la specificità, l’accuratezza ed i valori predittivi positivi e negativi erano del 91%.ConclusioniLa tecnica di analisi semiautomatica utilizzata appare accurata nell’identificazione di tessuto neoplastico e nell’elaborazione di mappe parametriche perfusionali, con il valore aggiunto di una valida misura dei parametri perfusionali sulle mappe colorimetriche stesse.

Prospective evaluation of drug-induced lung toxicity with high-resolution CT and transbronchial biopsy

PurposeThis study compared the results of high-resolution computed tomography (HRCT) and cytohistology after transbronchial biopsy in the evaluation of drug-related interstitial lung disease (DR-ILD).Materials and methodsPatients with a clinical and imaging diagnosis of DR-ILD were prospectively included in a study protocol lasting 5 years. All patients were evaluated by bronchoscopy with transbronchial biopsy or bronchoalveolar lavage (BAL) following an HRCT examination that raised a suspicion of DR-ILD. Two radiologists (one senior and one junior), unaware of the diagnosis, reported the single HRCT findings, their distribution and predominant pattern. In the event of disagreement, the diagnosis was subsequently reached by consensus. Cytohistological examination was considered the gold standard in the diagnosis of DR-ILD. Patients who were unable to undergo the endoscopic procedure were excluded from the study.ResultsThe study included 42 patients (25 men, 17 women; age range 20–84 years). Transbronchial biopsy was performed in all but four patients (one case of alveolar haemorrhage and three cases of lipoid pneumonia) in whom the diagnosis was established with BAL. Assessment of the HRCT images revealed the following patterns: noncardiogenic pulmonary oedema (n=13); organising pneumonia (OP) (n=9); hypersensitivity pneumonitis (HP) (n=2); alveolar haemorrhage (AH) (n=2); nonspecific interstitial pneumonia (NSIP) (n=5); lipoid pneumonia (LP) (n=1); sarcoid-like pattern (n=1). Cytohistological diagnosis revealed diffuse alveolar damage (DAD) in 11 patients, OP in seven, HP in three, AH in three, chronic interstitial pneumonia (CIP) in eight, LP in three and pseudosarcoidosis in one. Subdivision of the drugs into antineoplastic and nonantineoplastic agents showed that the most common patterns were CIP (n=6), DAD (n=2) and OP (n=2) in the antineoplastic group and DAD (n=9) and OP (n=5) in the nonantineoplastic group. Sensitivity and specificity of the radiological analysis was excellent, especially for patterns such as OP and DAD (sensitivity 0.86 and specificity 0.88 for OP; sensitivity 1 and specificity 0.93 for DAD).ConclusionsHRCT demonstrated excellent sensitivity and specificity. In cases in which its specificity was low, HRCT was nonetheless useful for biopsy planning and clinical-radiological monitoring after discontinuation of the drug treatment.RiassuntoObiettivoLo scopo del nostro lavoro è stato confrontare i reperti della tomografia computerizzata (TC) ad alta risoluzione (HRCT) con il dato cito-istologico fornito dalla biopsia transbronchiale, nella valutazione della tossicità polmonare farmaco indotta (DR-ILD).Materiali e metodiI pazienti con diagnosi clinicoradiologica di polmonite da farmaci sono stati prospettivamente inclusi in un protocollo di studio durato 5 anni. Tutti i pazienti erano sottoposti a broncoscopia, con biopsia transbronchiale o broncolavaggio (BAL) in seguito ad una HRCT che poneva il sospetto di DR-ILD. Due radiologi (un senior ed un junior) hanno riportato in cieco i dati riguardanti i singoli reperti, la distribuzione e il pattern predominante alla HRCT. In un secondo step, nei casi di rilevata discrepanza, la diagnosi è stata raggiunta per consenso. È stato considerato quale goal standard nel processo diagnostico di DR-ILD la diagnosi citoistologica. I pazienti che non potevano essere sottoposti a procedura endoscopica venivano esclusi dallo studio.RisultatiQuarantadue pazienti sono stati inclusi nello studio (25 maschi; 17 femmine; range di età 20–84 anni). La biopsia transbronchiale è stata effettuata in tutti i casi, fatta eccezione per un caso di emorragia alveolare e tre casi di polmonite lipoidea, in cui la diagnosi è stata raggiunta con il BAL. Le interpretazioni TC includevano i seguenti pattern: edema polmonare non cardiogeno (n=13); organizing pneumonia (OP) (n=9); hypersensitivity pneumonitis (HP) (n=2); alveolar haemorrage (AH) (n=2); non specific interstitial pneumonia (NSIP) (n=5); lipoid pneumonia (LP) (n=1); sarcoid-like pattern (n=1). La diagnosi anatomopatologica ha mostrato: danno alveolare diffuso (DAD) in 11 pazienti, OP in 7, HP in 3, AH in 3, polmonite interstiziale cronica (CIP) in 8, LP in 3; pseudosarcoidosi in 1. Dividendo i farmaci in antineoplastici e nonantineoplastici, i pattern più frequenti sono stati CIP (n=6); DAD (n=2); OP (n=2) per il gruppo degli antineoplastici. I pattern predominanti dei nonantineoplastici sono stati: DAD (n=9) e OP (n=5). Eccellente è stata la sensibilità e la specificità dell’analisi radiologica, soprattutto per pattern quali OP e DAD (OP: sensibilità pari a 0,86; specificità pari a 0,88; DAD: sensibilità 1; specificità 0,93).ConclusioniLa TC ad alta risoluzione ha mostrato un’eccellente sensibilità e specificità. Nei casi in cui la HRCT ha mostrato bassa specificità, è stata utile nel planning di una corretta biopsia e nel monitoraggio clnico-radiologico dopo la sospensione del farmaco.