Introduction. Exercise-Induced Laryngeal Obstruction (EILO) is a pathology that induces dyspnoea in affected subjects when they exercise at high workloads. EILO consists of the closure of the glottis, the median tract of larynx where vocal cords are positioned. The severity of EILO depends on the level of laryngeal closure. Moreover, laryngeal obstruction can be of two types, glottic or supraglottic. The latter is verified when supraglottic structures (i.e: cuneiform tubercles and aryepiglottic folds) collapse provoking the obstruction, the former happens when vocal cords adduct. In this case EILO is also called Vocal Cord Dysfunction (VCD). Glottic and supraglottic EILO can occur one independently from the other, or they can coexist. While for glottic obstruction there is not yet a validate treatment, supraglottic closure can be cured through laser ablation of cartilagineal structures. EILO can be easily misdiagnosed with other pathologies, mainly with Exercise-Induced Bronchoconstriction (EIB). Although they both show similar symptoms, they present important differences. EILO affects inspiration, it appears at maximum workload of physical exercise and it stops as soon as the effort finishes. EIB affects expiration and it appears as soon as the exercise stops. Often, EILO patients are initially treated with the pharmacologic therapy of EIB but without success. Failure of treatments is usually the first hint that leads to EILO diagnosis. Patients are found to suffer from EILO after 4.5 years on average from the onset of respiratory symptoms because of misdiagnosis. Until now, the most complete method to diagnose EILO is Continuous Laryngoscopy Exercise (CLE) test. A fiberscope is inserted through the nostril in the space between the epiglottis and the base of the tongue. Patient is therefore asked to perform physical effort until reaching the maximum limit. In the meanwhile, an expert laryngologist watches the video provided by the fiberscope on a monitor and gives scores from zero (normal closure) to three (severe obstruction) to the laryngeal closure, evaluating both glottic and supraglottic situations separately. This kind of assessment introduces subjectivity, even if in many studies laryngologists agreed on assigned scores. A more objective diagnosis can be done by EILOMEA, a software able to quantify different grades of obstruction. The operator is asked to select some points considered anatomically significant. EILOMEA evaluate only the singular figures chosen by the operator. It does not provide a dynamical assessment of glottic geometry. Little is known about the effects of EILO on ventilatory pattern and operational chest wall volumes. Aims and hypothesis. The aims of the present study are: 1) to implement a software able to dynamically and semi-automatically quantify the geometry of the larynx in terms of area and angle between the vocal cords; 2) to test the software on a population of healthy teenagers at rest, during vital capacity and during incremental exercise; 3) to verify if healthy male and female teenagers are characterized by different glottic geometry, ventilatory pattern and operational chest wall volumes assessed by opto-electronic plethysmography; 4) to verify if the software is able to detect an altered glottic pattern in EILO teenagers; and 5) to verify if EILO teenagers are characterized by different glottic geometry, ventilatory pattern and operational chest wall volumes compared to healthy peers. The main hypothesis is that EILO teenagers may alter their inspiratory timing to cope with the limitation induced by the exercise and this may influence their operational chest wall volumes. As the inspiratory reserve volume is mainly located in the ribcage, it is expected that ribcage operational volumes to be more effected by the presence of EILO. Materials and Methods. Forty-one teenagers (age: 15.7±0.5 years) attending the same scholastic year of a high school in Uppsala, Sweden, were analysed. To avoid misdiagnosis with EIB, they previously underwent to an asthma test to exclude this disease or to treat it. Successively, they were tested through a maximal, incremental, stepwise exercise on a cycle ergometer. After a period of quiet breathing (QB) for three minutes, during which they were asked to perform a vital capacity (VC), they began to pedal at a workload of 80 Watt and load was incremented of 40 Watt every two minutes, until they reached their maximum effort. During QB and the incremental exercise, simultaneous measurements of glottic video, through the fiberscope, and chest wall volumes, through Opto-Electronic Plethysmography (OEP) were performed. In synthesis, OEP is a system that computes chest wall volumes without requiring the use of a mouthpiece or a nose clip that are not compatible with the presence of the fiberscope. OEP is a motion analyser system based on infrared TV cameras and passive reflecting markers put on specific anatomical points. The system provides the 3D coordinates of each markers and through triangulation these points are linked to compute the enclosed volumes. Volumes of chest wall and of its three compartments, pulmonary and abdominal rib cage and abdomen, are measured. For every breath, it possible to compute tidal volume, as the difference between chest wall volume at the end of inspiration and end of expiration. Moreover, differences between end inspiratory and end expiratory volumes of each compartments were computed. Minute ventilation is the product of tidal volume and breathing frequency. The latter is the inverse of total breathing cycle duration. The ratio between inspiratory time and total breathing cycle duration (duty cycle) and the ratio between inspiratory and expiratory time were also computed. Tidal volume was also expressed as percentage of the vital capacity. A dedicated software was implemented to dynamically and semi-automatically quantify the geometry of the glottis. The software computes the area of the glottis opening and the angle formed by the two vocal cords, taking as input video from the fiberscope. The user is firstly asked to crop the image. Then the image is automatically sharpened and converted into a grey level image. Image is then processed through an histogram equalization to increase contrast. Furthermore, an average filter is applied to remove noise. The user is then asked to adjust contrast through a slider so that the intensity values of the image are remapped into new values so that a percentage of data is saturated at low and high intensities. This increases the contrast of the new image. After that, the operator sets a threshold to binarize the grey levels into a black and white image. The image has to be well trimmed. Finally the user selects the desired structure, and if it is not uniform inside, it is automatically filled with black. The result of this process is a black glottis on a white background. The software, then, automatically applies the same transformations to the following frames. If the image changes due to variations in luminosity, the parameters can be reset by the user. Starting from the black and white image, the area of the glottis is quantified by counting the black pixels. A linear interpolation of the points along the two vocal cords is then performed. The angle formed by the vocal cords is therefore computed as the difference between the angle of inclination of each line. The implemented software was validated by using another software, ImageJ, which allows the manual segmentation of the image. The three points at the extreme of the vocal cords were selected to compute the angle between them, while the perimeter of the glottis to compute the area. Results. Two out of the 41 subjects dropped out because they felt sick after the introduction of the fiberscope. The complete analysis was therefore performed on 39 teenagers, 10 of which were males and 29 females. Six subjects (all females) were diagnosed with EILO by the laryngologist watching the video provided by the fiberscope during exercise. They were judged as “moderate” patients (score>2) either glottic or supraglottic EILO. One of them was retested twice. The entire population was therefore split into 3 groups: healthy males, healthy females and pathological females. Results were compared between healthy females and healthy males, to verify if differences subsist because of the gender, and between healthy females and pathological females to verify the effect of EILO. The maximal workloads reached were 264±28 W, 191±31.1 W and 206.7±30.11 W respectively for healthy males, healthy females and pathological females. Workload was significantly higher in healthy males whereas no differences were found between healthy females and pathological females. Both geometrical glottic analysis and OEP data were performed during QB, and at the end 25%, 50%, 75% and 100% of maximum workload (WMax). Percentages were used to compare in a more uniform way the performance, considering that males reached higher loads. The validation of the implemented semi-automatic software, considering the manual segmentation as the gold standard, shows good agreement in both area and angle with percent errors of -0.57% and -7.2%, respectively. The implemented software therefore underestimates the angle computation that may be due to the linear interpolation of vocal cords, whereas glottis has a more complex shape. Vocal cords angles at QB and WMax are respectively 37.7°±9.6° and 42.9°±9.3° in healthy females, 28.8°±5.1° and 41.6°±5.6° in healthy males and 34.6°±15.9° and 36.2°±8° in pathological females. During QB, the angle is significantly lower in healthy males compared to healthy females. Area and angles were firstly evaluated during QB and VC. The forced inspiration and the forced expiration were considered separately. Changes in the values of the area were found. The area is higher at the end of maximal inspiration and lower at the end of maximal expiration with QB being intermediate between the two. Area and angles were then evaluated during QB, at 25%, 50%, 75% and 100% of WMAX. Area was normalized with the value at quiet breathing. While in healthy subjects the area immediately increases with the onset of exercise and it remains higher than rest until WMax, in the pathological population it increases until the 50% of the workload, but then it returns to QB values. As a consequence, the values of area at 75% and WMax of pathological females were lower than healthy females. Furthermore, the vocal cords angle in healthy subjects significantly increases at WMax compared to QB, while no changes are found in EILO patients. Healthy males reach higher values of ventilation and higher end inspiratory volumes in both the abdominal rib cage and the abdomen with respect to females. No other differences were found between healthy males and females. Healthy and pathological females, on the other hand, show significant difference in both duty cycle and the ratio between inspiratory and expiratory time at WMax. In pathological females, inspiratory time increases more respect expiratory time. Tidal volume, breathing frequency and minute ventilation are similar between the two female groups. Among the total and compartmental operational volumes, the only difference found is related to the pulmonary ribcage. The end-expiratory volume of the pulmonary ribcage becomes higher than QB at WMax. The end-inspiratory volume of the pulmonary ribcage becomes higher than healthy females at high level of exercise until WMax. Lastly, the repeatability of both measurements was evaluated comparing acquisitions of the subject who was analysed twice. Glottic area, vocal cords angle, ventilatory pattern and operational volumes follow the same trend with similar values. Conclusions. The incidence of EILO found is 15.4%, higher than what reported (<10%). A software able to dynamically and semi-automatically process the images from the fiberscope has been developed. It shows good reliability and all the results obtained are in line with those presented in the literature and visually evaluated by laryngologists: 1) the vocal cords angle at rest is lower in males. Males are known to be characterized by bigger laryngeal size. The combination of this increased structure and the lower vocal cord angle forms the protuberance called Adam’s apple; 2) glottic area increases at full inspiration, whereas it decreases at full expiration; 3) in healthy subjects the area of the glottis increases during exercise; and 4) in EILO females, the glottic geometry is limited only at high workloads. Healthy males reach higher maximal workload with greater ventilation due to increased tidal volume, compared with healthy females. With increasing exercise, the two groups of healthy subjects show similar pattern in terms of timing and operational volumes. The maximal workload and the ventilatory pattern are similar between EILO and healthy female. Instead, EILO females show a characteristic pattern of timing and operational volumes during exercise. At maximal exercise, therefore after the onset of EILO, they increase the inspiratory time and they shift the pulmonary ribcage operational volumes towards higher volumes. Both end-expiratory and end-inspiratory pulmonary rib cage volumes are higher than healthy peers, with similar expansion. In this way tidal volume is not limited in EILO patients. The increased inspiratory time and the consequent relative shortness of expiratory time may be the reason of the different pattern found in operational volumes. The action of the inspiratory muscles is therefore prolonged compared to the expiratory muscles. The fact that hyperinflation involves only pulmonary rib cage may be explained by the fact that it is stiffer than the other compartments. This is particularly evident during expiration. In fact, the expiratory reserve volume is almost completely located in the abdomen that is more compliant. The limits of the study concern the implemented software. Measurements of glottic sizes are in pixel, and it was not possible to convert them in metric system. Lens distortion, here not corrected, introduced errors in measurements. The computational time is very high and for this reason the videos were not entirely analysed. Future development could optimize the software in terms of: 1) vocal cords angle measurement by using higher interpolation order; 2) correction of the distortions; 3) reduction computational time; 4) possibility to convert pixel results into metric system, and 5) test a higher number of EILO patients. It would be interesting to see if the observed pattern of operational volumes is present also in moderate EILO males and more severe EILO patients.
Introduzione. L’Ostruzione Laringale Indotta dall’Esercizio (EILO), è una patologia che si manifesta ad alti carichi di esercizio inducendo dispnea nei soggetti affetti. EILO consiste nella chiusura della glottide, il tratto mediano della laringe dove sono posizionate le corde vocali. La severità di EILO dipende dal livello della chiusura laringale, che può essere di due tipi, glottica o supraglottica. La seconda si verifica con il collasso delle strutture supraglottiche (i.e. tubercoli cuneiformi e pieghe ariepiglottiche), la prima con l’adduzione delle corde vocali. In quest’ultimo caso EILO viene definita anche Disfunzione delle Corde Vocali (VCD). EILO glottica e supraglottica possono presentarsi indipendentemente l’una dall’altra o possono coesistere. Mentre per l’ostruzione glottica non esiste ancora un trattamento validato, la chiusura supraglottica può essere curata tramite ablazione laser delle strutture cartilaginee. La diagnosi di EILO può essere confusa con altre patologie, principalmente con la Broncocostrizione Indotta da Esercizio (EIB). Pur manifestando sintomi simili, presentano delle differenze. EILO si manifesta durante l’inspirazione, durante sforzo massimo e i sintomi scompaiono al termine dell’esercizio. EIB riguarda invece l’espirazione e si manifesta non appena lo sforzo termina. Solitamente, all’inizio, i pazienti EILO vengono curati con le terapie farmacologiche usate per EIB ma senza successo. Il fallimento nella cura è il primo indizio che porta alla diagnosi di EILO, la quale richiede in media 4.5 anni dal primo manifestarsi dei sintomi, poiché spesso vengono effettuate diagnosi errate. Fino ad ora, il metodo più completo per diagnosticare EILO è il test di Laringoscopia Continua durante Esercizio (CLE). Un fibroscopio viene inserito dalla narice fino allo spazio fra l’epiglottide e la base della lingua. Il paziente viene dunque sottoposto a esercizio fisico fino al raggiungimento del massimo sforzo. Un laringoiatra esperto, osservando il video registrato dal fibroscopio, assegna un punteggio da zero (chiusura normale) a tre (ostruzione severa) alla chiusura della laringe, valutando separatamente la situazione glottica e supraglottica. Questo tipo di diagnosi introduce soggettività, anche se in molti studi gli esperti laringoiatri concordano sui punteggi assegnati. Una diagnosi più oggettiva può essere stabilita da EILOMEA, un software capace di distinguere diversi gradi di ostruzione. Esso valuta le singole figure scelte dall’operatore senza analizzare dinamicamente la geometria glottica. L’operatore ha il compito di selezionare i punti anatomicamente significativi. Gli effetti di EILO sul pattern ventilatorio e sui volumi operazionali della parete toraco-addominale sono poco conosciuti. Scopi e ipotesi. Gli scopi del presente studio sono: 1) implementare un software capace di quantificare dinamicamente e semi-automaticamente la geometria della laringe in termini di area e angolo fra le corde vocali; 2) di testare il software su una popolazione di adolescenti a riposo, durante la capacità vitale e durante esercizio incrementale; 3) di verificare se adolescenti maschi e femmine sani sono caratterizzati da differenze nella geometria glottica, nel pattern ventilatorio e nei volumi operazionali della parete toraco-addominale misurati da pletismografia optoelettronica; 4) di verificare se il software è capace di individuare un pattern glottico alterato negli adolescenti affetti da EILO; 5) di verificare se gli adolescenti affetti da EILO sono caratterizzati da diversa geometria glottica, diverso pattern ventilatorio e diversi volumi operazionali toraco-addominali rispetto ai sani. L’ipotesi principale è che gli adolescenti affetti da EILO possano alterare le tempistiche inspiratorie per gestire la limitazione indotta dall’esercizio e questo possa influenzare i volumi operazionali toraco-addominali. Si sospetta che i volumi operazionali della gabbia toracica siano i più colpiti dalla presenza di EILO Materiali e Metodi. Sono stati analizzati quarantuno adolescenti (età: 15.7±0.5 anni) frequentanti lo stesso anno scolastico di una scuola superiore in Uppsala, Svezia. Per evitare diagnosi errate, essi sono stati sottoposti ad un test per escludere la presenza di EIB o eventualmente curarla. Successivamente, i soggetti sono stati testati tramite un esercizio massimale, incrementale, a gradino, su un cicloergometro. Dopo un periodo di respiro spontaneo (QB) di tre minuti, nel quale viene chiesto ai soggetti di eseguire una manovra di capacità vitale (VC), hanno iniziato a pedale ad un carico di 80 Watt e il carico è stato incrementato di 40 Watt ogni due minuti fino al raggiungimento del massimo sforzo. Sia durante QB sia durante esercizio, sono state effettuate misure di volumi della parete toraco-addominale con Pletismografia Opto-Elettronica (OEP) ed un video della glottide è stato registrato tramite fibroscopio. In sintesi, OEP è un sistema che misura volumi toraco-addominali senza richiedere l’uso di strumentazione in bocca o nel naso, non interferendo quindi con la presenza del fibroscopio. OEP è un sistema che analizza il movimento basandosi su telecamere ad infrarossi e marcatori passivi riflettenti che vengono posizionati in specifici punti anatomici. Il sistema individua le coordinate 3D di ogni marcatore e misura i volumi che vengono individuati dal collegamento dei punti ottenuto tramite triangolazione. Sono stati misurati i volumi della parete toraco-addominale e dei suoi tre compartimenti: la gabbia toracica polmonare e addominale e l’addome. Per ogni respiro, è possibile calcolare il volume corrente come differenza del volume della parete toraco-addominale a fine inspirazione e a fine espirazione. Inoltre, sono stati misurati i valori differenziali fra fine inspirazione e fine espirazione di ogni singolo compartimento. La ventilazione al minuto si ottiene dal prodotto fra il volume corrente e la frequenza respiratoria. Quest’ultima è data dall’inverso della durata del ciclo respiratorio. Sono stati misurati inoltre il rapporto fra tempo inspiratorio e la durata totale del ciclo respiratorio (duty cycle) e il rapporto fra tempo inspiratorio ed espiratorio. Il volume corrente è stato valutato anche come percentuale della capacità vitale. La geometria della glottide è stata quantificata in modo dinamico e semi-automatico attraverso un software appositamente implementato. Il software misura l’area dell’apertura glottica e l’angolo fra le corde vocali, prendendo come input il video dal fibroscopio. Inizialmente, all’operatore è chiesto di ritagliare l’immagine. Successivamente l’immagine viene automaticamente messa a fuoco e convertita nei livelli di grigio. Poi, l’istogramma dell’immagine viene equalizzato per aumentare il contrasto. Un filtro media viene utilizzato per rimuovere il rumore. Inizialmente l’operatore aggiusta il contrasto tramite un cursore, saturando una certa percentuale di intensità basse e alte, mentre le restanti vengono ri-mappate in nuovi valori di intensità, aumentando il contrasto. L’operatore sceglie inoltre una soglia che suddivida i livelli di grigio in bianco e nero, in modo che l’immagine sia ben contornata. Quindi, l’operatore seleziona la struttura desiderata, che viene riempita qualora non fosse uniforme al suo interno. L’immagine della glottide risulta infine nera su uno sfondo bianco. Il software applica automaticamente le stesse trasformazioni alle immagini successive. Se variazioni nelle intensità cambiano la visualizzazione dell’immagine, i parametri possono essere resettati dall’operatore. Partendo dall’immagine in bianco e nero, l’area della glottide viene misurata contando i pixel neri. I punti lungo le due corde vocali vengono interpolati linearmente. L’angolo formato dalle corde vocali è quindi calcolato come differenza fra gli angoli di inclinazione delle due linee. Il software implementato è stato validato utilizzando un altro software, ImageJ, che permette la segmentazione manuale dell’immagine. L’operatore seleziona tre punti agli estremi delle corde vocali per calcolarne l’angolo all’interno, mentre l’area viene misurata tramite la selezione del perimetro. Risultati. Due soggetti su 41 non hanno partecipato allo studio poiché si sono sentiti male all’introduzione del fibroscopio. L’analisi completa è stata effettuata su 39 adolescenti, 10 dei quali maschi e 29 femmine. EILO è stata diagnosticata a sei soggetti (tutte femmine) dal laringoiatra guardando il video registrato dal fibroscopio durante l’esercizio. Sono stati classificati come pazienti “moderati” (punteggio>2) con EILO glottica o supraglottica. Un paziente è stato ri-testato. I soggetti sono stati divisi in tre popolazioni: maschi sani, femmine sane e femmine patologiche. I risultati sono stati comparati fra femmine sani e maschi sani, per verificare qualora ci fossero differenze fra i generi e fra femmine sane e femmine patologiche per studiare gli effetti di EILO. I massimi carichi raggiunti sono stati: 264±28 W, 191±31.1 W e 206.7±30.11 W rispettivamente per maschi sani, femmine sane e femmine patologiche. Il carico raggiunto è significativamente alto nei maschi sani, mentre non sono state trovate differenze fra femmine sane e patologiche. L’analisi della geometria glottica e dei dati misurati tramite OEP sono state effettuate sia durante QB sia a fine del 25%, 50%, 75% e 100% del carico massimo (WMax). In questo modo le performance sono state considerate in modo più uniforme, a maggior ragione data la differenza nel carico raggiunto da maschi e femmine sane. La validazione del software semi-automatico implementato, considerando come gold standard la segmentazione manuale, mostra buoni risultati con errori percentuali di -0.57% e -7.2% rispettivamente nella misura di area e angolo. La sottostima dell’angolo da parte del software implementato è probabilmente dovuta all’interpolazione lineare delle corde vocali, mentre la glottide presenta una forma più complessa. Gli angoli fra le corde vocali a QB e WMax sono rispettivamente 37.7°±9.6° e 42.9°±9.3° nelle femmine sane, 28.8°±5.1° e 41.6°±5.6° nei maschi sani e 34.6°±15.9° e 36.2°±8° nelle femmine patologiche. Durante QB, l’angolo è significativamente più basso nei maschi sani rispetto alle femmine sane. Area e angolo sono stati inizialmente valutati durante QB e VC. Inspirazione ed espirazione forzata sono state considerate separatamente. L’area è stata misurata come più grossa alla fine della massima inspirazione e più piccola alla fine della massima espirazione, con il QB situato nel mezzo. Area e angolo sono stati poi valutati durante QB, al 25%, 50%, 75% e 100% di WMAX. L’area è stata considerata in funzione del suo valore a QB. Mentre nei soggetti sani l’area aumenta non appena il soggetto inizia l’esercizio e rimane maggiore dei valori a riposo fino a WMax, nei soggetti patologici aumenta fino al 50% del carico massimale, ritornando poi a valori simili al QB. Di conseguenza, i valori dell’area al 75% e a WMax delle femmine patologiche sono minori rispetto a quelli delle femmine sane. In aggiunta, l’angolo fra le corde vocali nei soggetti sani aumenta significativamente a WMax rispetto a QB, mentre nessuna differenza si riscontra nei soggetti EILO. I maschi sani raggiungono maggiori valori di ventilazione e maggiori volumi a fine inspirazione sia nella gabbia toracica addominale sia nell’addome rispetto alle femmine. Non sono state trovate altre rilevanti differenze fra i maschi e le femmine sane. Femmine sane e patologiche, invece, mostrano diversi valori di duty cycle e di rapporto fra tempo inspiratorio ed espiratorio a WMax. Nelle femmine patologiche, il tempo inspiratorio aumenta di più rispetto a quello espiratorio. Valori simili di volume corrente, frequenza respiratoria e ventilazione al minuto sono stati misurati nelle due popolazioni. L’unica differenza fra i volumi operazionali totali e compartimentali è data dalla gabbia toracica polmonare. Il volume di fine espirazione della gabbia toracica polmonare aumenta rispetto a QB a WMax. Il volume di fine inspirazione della gabbia toracica polmonare aumenta rispetto a QB ad alti carichi fino a WMax. Infine, la ripetibilità di entrambe le misurazioni è stata valutata comparando le acquisizioni del soggetto che è stato ri-testato. L’area glottica, l’angolo fra le corde vocali, l’andamento della ventilazione e i volumi operazionali seguono lo stesso andamento con valori simili. Conclusioni. L’incidenza di EILO in questo studio è stata del 15.4%, più alta di quanto riportato in letteratura (<10%). E’ stato implementato un software capace di processare dinamicamente e semi-automaticamente le immagini del fibroscopio. Risulta un metodo di misura affidabile e tutti i risultati sono in linea con quelli presenti in letteratura e valutati visivamente dai laringoiatri: 1) l’angolo fra le corde vocali a riposo è minore nei maschi. I maschi sono caratterizzati da una struttura laringale più grande delle femmine. La combinazione di una struttura maggiore e un angolo fra le corde vocali minore forma la protuberanza chiamata pomo d’Adamo; 2) l’area glottica aumenta a fine inspirazione e diminuisce a fine espirazione; 3) nei soggetti sani l’area della glottide aumenta durante esercizio; e 4) nelle femmine EILO, la geometria glottica è limitata solo a carichi alti. I maschi sani raggiungono carichi di esercizio più alti combinati ad una più alta ventilazione dovuta a un volume corrente maggiore rispetto alle femmine. I due gruppi di soggetti sani mostrano andamenti simili dei volumi operazionali e delle tempistiche di respirazione all’aumentare dell’esercizio. Il carico massimo e il pattern ventilatorio sono simili nelle femmine sane e patologiche. Al contrario, le femmine EILO mostrano un andamento diverso nelle tempistiche respiratorie e nei volumi operazionali durante esercizio. A massimo esercizio, mentre EILO si manifesta, aumentano il tempo inspiratorio e traslano i volumi operazionali della gabbia toracica polmonare verso valori più alti. Entrambi i valori di volume della gabbia toracica a fine inspirazione e fine espirazione sono più alti rispetto a quelli delle femmine sane, ma con espansioni comparabili. Il volume corrente non è dunque limitato nei pazienti EILO. Il diverso andamento dei volumi operazionali potrebbe essere dato dall’aumento nel tempo inspiratorio e dalla conseguente riduzione del tempo espiratorio. L’azione dei muscoli inspiratori è quindi prolungata rispetto a quella dei muscoli espiratori. Il coinvolgimento della sola gabbia toracica polmonare nell’insufflazione potrebbe essere causato dalla sua rigidezza rispetto agli altri compartimenti, con particolare evidenza nell’espirazione. Infatti, il volume di riserva espiratoria è maggiormente allocato nell’addome, che è più compliante. I limiti dello studio riguardano il software implementato. Le misurazioni delle dimensioni glottiche sono in pixel, e non è stato possibile convertirle nel sistema metrico. La distorsione della lente, non corretta, introduce degli errori nelle misurazioni. Il tempo computazionale del software è molto alto e per questa ragione i video non sono stati analizzati interamente. Sviluppi futuri potrebbero ottimizzare il software in termini di: 1) misura dell’angolo fra le corde vocali usando interpolazioni di ordine più alto; 2) correzione della distorsione della lente; 3) riduzione del tempo computazionale; 4) possibilità di convertire i risultati in pixel in misure del sistema metrico, e 5) testare un maggior numero di pazienti EILO. Sarebbe interessante vedere se l’andamento osservato nei volumi operazionali è presente anche in maschi aventi EILO moderata e in casi di EILO più gravi.
Dynamical changes of glottic geometry and chest wall volumes during exercise induced laryngeal obstruction (EILO)
MARIANI, FRANCESCA
2015/2016
Abstract
Introduction. Exercise-Induced Laryngeal Obstruction (EILO) is a pathology that induces dyspnoea in affected subjects when they exercise at high workloads. EILO consists of the closure of the glottis, the median tract of larynx where vocal cords are positioned. The severity of EILO depends on the level of laryngeal closure. Moreover, laryngeal obstruction can be of two types, glottic or supraglottic. The latter is verified when supraglottic structures (i.e: cuneiform tubercles and aryepiglottic folds) collapse provoking the obstruction, the former happens when vocal cords adduct. In this case EILO is also called Vocal Cord Dysfunction (VCD). Glottic and supraglottic EILO can occur one independently from the other, or they can coexist. While for glottic obstruction there is not yet a validate treatment, supraglottic closure can be cured through laser ablation of cartilagineal structures. EILO can be easily misdiagnosed with other pathologies, mainly with Exercise-Induced Bronchoconstriction (EIB). Although they both show similar symptoms, they present important differences. EILO affects inspiration, it appears at maximum workload of physical exercise and it stops as soon as the effort finishes. EIB affects expiration and it appears as soon as the exercise stops. Often, EILO patients are initially treated with the pharmacologic therapy of EIB but without success. Failure of treatments is usually the first hint that leads to EILO diagnosis. Patients are found to suffer from EILO after 4.5 years on average from the onset of respiratory symptoms because of misdiagnosis. Until now, the most complete method to diagnose EILO is Continuous Laryngoscopy Exercise (CLE) test. A fiberscope is inserted through the nostril in the space between the epiglottis and the base of the tongue. Patient is therefore asked to perform physical effort until reaching the maximum limit. In the meanwhile, an expert laryngologist watches the video provided by the fiberscope on a monitor and gives scores from zero (normal closure) to three (severe obstruction) to the laryngeal closure, evaluating both glottic and supraglottic situations separately. This kind of assessment introduces subjectivity, even if in many studies laryngologists agreed on assigned scores. A more objective diagnosis can be done by EILOMEA, a software able to quantify different grades of obstruction. The operator is asked to select some points considered anatomically significant. EILOMEA evaluate only the singular figures chosen by the operator. It does not provide a dynamical assessment of glottic geometry. Little is known about the effects of EILO on ventilatory pattern and operational chest wall volumes. Aims and hypothesis. The aims of the present study are: 1) to implement a software able to dynamically and semi-automatically quantify the geometry of the larynx in terms of area and angle between the vocal cords; 2) to test the software on a population of healthy teenagers at rest, during vital capacity and during incremental exercise; 3) to verify if healthy male and female teenagers are characterized by different glottic geometry, ventilatory pattern and operational chest wall volumes assessed by opto-electronic plethysmography; 4) to verify if the software is able to detect an altered glottic pattern in EILO teenagers; and 5) to verify if EILO teenagers are characterized by different glottic geometry, ventilatory pattern and operational chest wall volumes compared to healthy peers. The main hypothesis is that EILO teenagers may alter their inspiratory timing to cope with the limitation induced by the exercise and this may influence their operational chest wall volumes. As the inspiratory reserve volume is mainly located in the ribcage, it is expected that ribcage operational volumes to be more effected by the presence of EILO. Materials and Methods. Forty-one teenagers (age: 15.7±0.5 years) attending the same scholastic year of a high school in Uppsala, Sweden, were analysed. To avoid misdiagnosis with EIB, they previously underwent to an asthma test to exclude this disease or to treat it. Successively, they were tested through a maximal, incremental, stepwise exercise on a cycle ergometer. After a period of quiet breathing (QB) for three minutes, during which they were asked to perform a vital capacity (VC), they began to pedal at a workload of 80 Watt and load was incremented of 40 Watt every two minutes, until they reached their maximum effort. During QB and the incremental exercise, simultaneous measurements of glottic video, through the fiberscope, and chest wall volumes, through Opto-Electronic Plethysmography (OEP) were performed. In synthesis, OEP is a system that computes chest wall volumes without requiring the use of a mouthpiece or a nose clip that are not compatible with the presence of the fiberscope. OEP is a motion analyser system based on infrared TV cameras and passive reflecting markers put on specific anatomical points. The system provides the 3D coordinates of each markers and through triangulation these points are linked to compute the enclosed volumes. Volumes of chest wall and of its three compartments, pulmonary and abdominal rib cage and abdomen, are measured. For every breath, it possible to compute tidal volume, as the difference between chest wall volume at the end of inspiration and end of expiration. Moreover, differences between end inspiratory and end expiratory volumes of each compartments were computed. Minute ventilation is the product of tidal volume and breathing frequency. The latter is the inverse of total breathing cycle duration. The ratio between inspiratory time and total breathing cycle duration (duty cycle) and the ratio between inspiratory and expiratory time were also computed. Tidal volume was also expressed as percentage of the vital capacity. A dedicated software was implemented to dynamically and semi-automatically quantify the geometry of the glottis. The software computes the area of the glottis opening and the angle formed by the two vocal cords, taking as input video from the fiberscope. The user is firstly asked to crop the image. Then the image is automatically sharpened and converted into a grey level image. Image is then processed through an histogram equalization to increase contrast. Furthermore, an average filter is applied to remove noise. The user is then asked to adjust contrast through a slider so that the intensity values of the image are remapped into new values so that a percentage of data is saturated at low and high intensities. This increases the contrast of the new image. After that, the operator sets a threshold to binarize the grey levels into a black and white image. The image has to be well trimmed. Finally the user selects the desired structure, and if it is not uniform inside, it is automatically filled with black. The result of this process is a black glottis on a white background. The software, then, automatically applies the same transformations to the following frames. If the image changes due to variations in luminosity, the parameters can be reset by the user. Starting from the black and white image, the area of the glottis is quantified by counting the black pixels. A linear interpolation of the points along the two vocal cords is then performed. The angle formed by the vocal cords is therefore computed as the difference between the angle of inclination of each line. The implemented software was validated by using another software, ImageJ, which allows the manual segmentation of the image. The three points at the extreme of the vocal cords were selected to compute the angle between them, while the perimeter of the glottis to compute the area. Results. Two out of the 41 subjects dropped out because they felt sick after the introduction of the fiberscope. The complete analysis was therefore performed on 39 teenagers, 10 of which were males and 29 females. Six subjects (all females) were diagnosed with EILO by the laryngologist watching the video provided by the fiberscope during exercise. They were judged as “moderate” patients (score>2) either glottic or supraglottic EILO. One of them was retested twice. The entire population was therefore split into 3 groups: healthy males, healthy females and pathological females. Results were compared between healthy females and healthy males, to verify if differences subsist because of the gender, and between healthy females and pathological females to verify the effect of EILO. The maximal workloads reached were 264±28 W, 191±31.1 W and 206.7±30.11 W respectively for healthy males, healthy females and pathological females. Workload was significantly higher in healthy males whereas no differences were found between healthy females and pathological females. Both geometrical glottic analysis and OEP data were performed during QB, and at the end 25%, 50%, 75% and 100% of maximum workload (WMax). Percentages were used to compare in a more uniform way the performance, considering that males reached higher loads. The validation of the implemented semi-automatic software, considering the manual segmentation as the gold standard, shows good agreement in both area and angle with percent errors of -0.57% and -7.2%, respectively. The implemented software therefore underestimates the angle computation that may be due to the linear interpolation of vocal cords, whereas glottis has a more complex shape. Vocal cords angles at QB and WMax are respectively 37.7°±9.6° and 42.9°±9.3° in healthy females, 28.8°±5.1° and 41.6°±5.6° in healthy males and 34.6°±15.9° and 36.2°±8° in pathological females. During QB, the angle is significantly lower in healthy males compared to healthy females. Area and angles were firstly evaluated during QB and VC. The forced inspiration and the forced expiration were considered separately. Changes in the values of the area were found. The area is higher at the end of maximal inspiration and lower at the end of maximal expiration with QB being intermediate between the two. Area and angles were then evaluated during QB, at 25%, 50%, 75% and 100% of WMAX. Area was normalized with the value at quiet breathing. While in healthy subjects the area immediately increases with the onset of exercise and it remains higher than rest until WMax, in the pathological population it increases until the 50% of the workload, but then it returns to QB values. As a consequence, the values of area at 75% and WMax of pathological females were lower than healthy females. Furthermore, the vocal cords angle in healthy subjects significantly increases at WMax compared to QB, while no changes are found in EILO patients. Healthy males reach higher values of ventilation and higher end inspiratory volumes in both the abdominal rib cage and the abdomen with respect to females. No other differences were found between healthy males and females. Healthy and pathological females, on the other hand, show significant difference in both duty cycle and the ratio between inspiratory and expiratory time at WMax. In pathological females, inspiratory time increases more respect expiratory time. Tidal volume, breathing frequency and minute ventilation are similar between the two female groups. Among the total and compartmental operational volumes, the only difference found is related to the pulmonary ribcage. The end-expiratory volume of the pulmonary ribcage becomes higher than QB at WMax. The end-inspiratory volume of the pulmonary ribcage becomes higher than healthy females at high level of exercise until WMax. Lastly, the repeatability of both measurements was evaluated comparing acquisitions of the subject who was analysed twice. Glottic area, vocal cords angle, ventilatory pattern and operational volumes follow the same trend with similar values. Conclusions. The incidence of EILO found is 15.4%, higher than what reported (<10%). A software able to dynamically and semi-automatically process the images from the fiberscope has been developed. It shows good reliability and all the results obtained are in line with those presented in the literature and visually evaluated by laryngologists: 1) the vocal cords angle at rest is lower in males. Males are known to be characterized by bigger laryngeal size. The combination of this increased structure and the lower vocal cord angle forms the protuberance called Adam’s apple; 2) glottic area increases at full inspiration, whereas it decreases at full expiration; 3) in healthy subjects the area of the glottis increases during exercise; and 4) in EILO females, the glottic geometry is limited only at high workloads. Healthy males reach higher maximal workload with greater ventilation due to increased tidal volume, compared with healthy females. With increasing exercise, the two groups of healthy subjects show similar pattern in terms of timing and operational volumes. The maximal workload and the ventilatory pattern are similar between EILO and healthy female. Instead, EILO females show a characteristic pattern of timing and operational volumes during exercise. At maximal exercise, therefore after the onset of EILO, they increase the inspiratory time and they shift the pulmonary ribcage operational volumes towards higher volumes. Both end-expiratory and end-inspiratory pulmonary rib cage volumes are higher than healthy peers, with similar expansion. In this way tidal volume is not limited in EILO patients. The increased inspiratory time and the consequent relative shortness of expiratory time may be the reason of the different pattern found in operational volumes. The action of the inspiratory muscles is therefore prolonged compared to the expiratory muscles. The fact that hyperinflation involves only pulmonary rib cage may be explained by the fact that it is stiffer than the other compartments. This is particularly evident during expiration. In fact, the expiratory reserve volume is almost completely located in the abdomen that is more compliant. The limits of the study concern the implemented software. Measurements of glottic sizes are in pixel, and it was not possible to convert them in metric system. Lens distortion, here not corrected, introduced errors in measurements. The computational time is very high and for this reason the videos were not entirely analysed. Future development could optimize the software in terms of: 1) vocal cords angle measurement by using higher interpolation order; 2) correction of the distortions; 3) reduction computational time; 4) possibility to convert pixel results into metric system, and 5) test a higher number of EILO patients. It would be interesting to see if the observed pattern of operational volumes is present also in moderate EILO males and more severe EILO patients.File | Dimensione | Formato | |
---|---|---|---|
TESIFRANCESCAMARIANI.pdf
accessibile in internet solo dagli utenti autorizzati
Descrizione: Testo della Tesi
Dimensione
4.76 MB
Formato
Adobe PDF
|
4.76 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/10589/133299