**4.1.4 Imaging features suggesting calcification**

Calcification of the GP has also been reported in the literature (Illum 1980; Lugaresi, Montagna et al. 1990; Adam, Baulac et al. 2008). The clinical presentations included acute neurological deficits with loss of initiative and slowness of thinking and acting (Adam, Baulac et al. 2008), and delayed neurological deficits with personality changes and akinesia (Lugaresi, Montagna et al. 1990). However one case was free of any neurological sequelae after 48 years of follow-up (Illum 1980).

### **4.1.5 Functional imaging features suggesting hypometabolism**

[18F]fluorodeoxyglucose (FDG) PET has been used to evaluate glucose metabolism activity. Decreased metabolism in the basal ganglion and frontal lobe has been frequently reported (Tengvar, Johansson et al. 2004; Hon, Yeung et al. 2006). The largest series on PET and CO intoxication with basal ganglion lesions included eight patients with their behavioral and MRI patterns (Laplane, Levasseur et al. 1989). Seven patients revealed hypometabolism of the prefrontal cortex in relation to other parts of the brain, leading to a concept of prefrontalpallidum circuit dysfunction. A functional study using [18F] F-DOPA showed presynaptic dopaminergic deficits in one case with parkinsonism symptoms after CO intoxication (Rissanen, Paavilainen et al. 2010). In this case, normal uptake of [11C] raclopride implicated normal postsynaptic dopaminergic function (Rissanen, Paavilainen et al. 2010).

Single photon emission computed tomography (SPECT) provides perfusion patterns of GM and the basal ganglion (Chang, Liu et al. 2008) with tracers such as 99mTc-ethylcysteinate dimer and 99mTc-Hexamethylpropyleneamine oxime. (99mTc-ECD) brain SPECT is considered to be more sensitive than brain CT for the early detection of hypoperfusion status (Wu, Changlai et al. 2003). In the acute stage, 50% to 85% of the patients with CO intoxication have been reported to have basal ganglion hypoperfusion (Wu, Changlai et al. 2003; Pach, Hubalewska et al. 2004).

Fig. 4. [18F]fluorodeoxyglucose positron emission tomography (PET) of two patients after CO poisoning.

Two and a half months after CO intoxication, a 33-year-old patient's CT showed low intensity of the globus pallidus (4A) on brain computed tomography (CT) while PET revealed a remarkably reduced uptake of FDG in bilateral striatum (arrows) and thalamus (4B). Five months after CO intoxication, another 36-year-old patient's CT showed no

Neuroimaging Studies in Carbon Monoxide Intoxication 361

WM area show low ADC values with high DWI intensities, while vasogenic edema shows

One month after CO intoxication, a 41-year-old woman with white matter hyperintensity in

The prevalence of imaging features suggesting WM demyelination or axonopathy range from 12% to 100% in CO intoxication (Chang, Han et al. 1992; Parkinson, Hopkins et al. 2002). The largest MRI study focusing on WM included 73 patients scanned on day 1, 2 weeks and 6 months after CO intoxication (Parkinson, Hopkins et al. 2002). Semiquantitative scores were rated on bilateral periventricular and centrum semiovale areas (Parkinson, Hopkins et al. 2002). Twelve percent of the patients had WM hyperintensities on T2WI on day 1 (Parkinson, Hopkins et al. 2002) with significantly more periventricular, but not centrum semiovale distributions as compared with age-matched controls. The WM lesions in the CO group did not change from day 1 to 6 months follow-up, however the hyperintensities in the centrum semiovale were related to worse cognitive performance. The study revealed no correlation between WM hyperintensities and carboxyhemoglobin level, or duration of CO exposure at any of the three scan times (Parkinson, Hopkins et al. 2002). Hyperintensities in T2WI and fluid-attenuated inversion recovery (FLAIR) and hypointensities in T1WI often suggest WM demyelination or axonopathy (Chang, Han et al. 1992; Pavese, Napolitano et al. 1999; Parkinson, Hopkins et al. 2002). From a pathological perspective, myelin damage is constant and can vary from discrete perivascular lesions to extensive periventricular demyelination and/or axonal destruction (Funata, Okeda et al. 1982; Prockop and Chichkova 2007). An autopsy study after CO intoxication showed that diffuse WM hyperintensities reflected apoptosis of oligodendrocytes (Akaiwa, Hozumi et al. 2002). Another autopsy study of brains three days after CO intoxication revealed a normal cortex and injured WM with disrupted myelin and pyknotic oligodendroglia, whilst the

DWI (6A) and iso- to low-signal intensity in ADC (6B) indicating cytotoxic edema.

**4.2.2 Imaging features suggesting WM demyelination or axonopathy** 

axons, astrocytes and capillaries were normal (Foncin and Le Beau 1978).

Fig. 6. A wide spectrum of white matter hyperintensities in fluid-attenuated inversion

recovery after carbon monoxide intoxication with cognitive deficits.

high signals on both sequences.

obvious lesions (4C, 4E) while PET revealed normal FDG uptake in bilateral striatum (4D, 4F arrows) and normal thalamic uptake.

### **4.1.6 Imaging features suggesting pallidoreticular damage**

In CO intoxication, pallidoreticular damage specifically targeting the fiber tract along the pallidum and substantia nigra pars reticulata was first described by Auer and Benveniste (Auer and Benveniste 1996). One case report revealed cytotoxic edema of bilateral GP with concurrent substantia nigra pars reticulata involvement in a patient scanned 12 days after CO intoxication (Kinoshita, Sugihara et al. 2005). Two case reports revealed pallidoreticular distribution after one year showing hyperintensities on T2WI and hypointensities on T1WI (Kawanami, Kato et al. 1998; Gandini, Prockop et al. 2002). The authors suggested that these two iron rich regions had selective tissue vulnerability due to the high affinity of CO to heme molecules (Kawanami, Kato et al. 1998; Gandini, Prockop et al. 2002; Kinoshita, Sugihara et al. 2005).

#### **4.2 WM lesions**

An increasing number of studies have established that WM lesions are the most common findings in CO intoxication patients, either in the acute phase or in those with delayed neuropsychiatric sequelae (Miura, Mitomo et al. 1985; Chang, Han et al. 1992; Choi, Kim et al. 1993; Lee and Marsden 1994). The largest study included 129 patients, and 33% of them had WM lesions on brain CT (Choi, Kim et al. 1993). In patients with improvements of neurological deficits, resolution of WM changes have also been noted (Klostermann, Vieregge et al. 1993; Matsushita, Takahashi et al. 1996; Pavese, Napolitano et al. 1999). Lesions of the WM area are believed to be associated with clinical outcomes (Miura, Mitomo et al. 1985; Vieregge, Klostermann et al. 1989; Choi, Kim et al. 1993).

#### **4.2.1 Imaging features suggesting WM cytotoxic/vasogenic edema**

In a pathological series, cytotoxic and vasogenic edema after CO intoxication were often mixed within three months, and the presence of cytotoxic edema was often noted to be in the acute phase (Ginsberg, Myers et al. 1974; Ginsberg 1985; Thom, Bhopale et al. 2004). The presence of cytotoxic edema lesions can be detected as early as the first day of CO intoxication (Sener 2003) or during the delayed phase (Murata, Kimura et al. 2001; Kim, Chang et al. 2003; Chu, Jung et al. 2004). Imaging features suggesting cytotoxic edema of the

Fig. 5. Diffusion weighted image (5A) and apparent diffusion coefficient (5B) in one case presenting as delayed neuropsychiatric sequelae after carbon monoxide intoxication.

obvious lesions (4C, 4E) while PET revealed normal FDG uptake in bilateral striatum (4D, 4F

In CO intoxication, pallidoreticular damage specifically targeting the fiber tract along the pallidum and substantia nigra pars reticulata was first described by Auer and Benveniste (Auer and Benveniste 1996). One case report revealed cytotoxic edema of bilateral GP with concurrent substantia nigra pars reticulata involvement in a patient scanned 12 days after CO intoxication (Kinoshita, Sugihara et al. 2005). Two case reports revealed pallidoreticular distribution after one year showing hyperintensities on T2WI and hypointensities on T1WI (Kawanami, Kato et al. 1998; Gandini, Prockop et al. 2002). The authors suggested that these two iron rich regions had selective tissue vulnerability due to the high affinity of CO to heme molecules (Kawanami, Kato et al. 1998; Gandini, Prockop et al. 2002; Kinoshita,

An increasing number of studies have established that WM lesions are the most common findings in CO intoxication patients, either in the acute phase or in those with delayed neuropsychiatric sequelae (Miura, Mitomo et al. 1985; Chang, Han et al. 1992; Choi, Kim et al. 1993; Lee and Marsden 1994). The largest study included 129 patients, and 33% of them had WM lesions on brain CT (Choi, Kim et al. 1993). In patients with improvements of neurological deficits, resolution of WM changes have also been noted (Klostermann, Vieregge et al. 1993; Matsushita, Takahashi et al. 1996; Pavese, Napolitano et al. 1999). Lesions of the WM area are believed to be associated with clinical outcomes (Miura, Mitomo

In a pathological series, cytotoxic and vasogenic edema after CO intoxication were often mixed within three months, and the presence of cytotoxic edema was often noted to be in the acute phase (Ginsberg, Myers et al. 1974; Ginsberg 1985; Thom, Bhopale et al. 2004). The presence of cytotoxic edema lesions can be detected as early as the first day of CO intoxication (Sener 2003) or during the delayed phase (Murata, Kimura et al. 2001; Kim, Chang et al. 2003; Chu, Jung et al. 2004). Imaging features suggesting cytotoxic edema of the

Fig. 5. Diffusion weighted image (5A) and apparent diffusion coefficient (5B) in one case presenting as delayed neuropsychiatric sequelae after carbon monoxide intoxication.

arrows) and normal thalamic uptake.

Sugihara et al. 2005).

**4.2 WM lesions** 

**4.1.6 Imaging features suggesting pallidoreticular damage** 

et al. 1985; Vieregge, Klostermann et al. 1989; Choi, Kim et al. 1993).

**4.2.1 Imaging features suggesting WM cytotoxic/vasogenic edema** 

WM area show low ADC values with high DWI intensities, while vasogenic edema shows high signals on both sequences.

One month after CO intoxication, a 41-year-old woman with white matter hyperintensity in DWI (6A) and iso- to low-signal intensity in ADC (6B) indicating cytotoxic edema.

#### **4.2.2 Imaging features suggesting WM demyelination or axonopathy**

The prevalence of imaging features suggesting WM demyelination or axonopathy range from 12% to 100% in CO intoxication (Chang, Han et al. 1992; Parkinson, Hopkins et al. 2002). The largest MRI study focusing on WM included 73 patients scanned on day 1, 2 weeks and 6 months after CO intoxication (Parkinson, Hopkins et al. 2002). Semiquantitative scores were rated on bilateral periventricular and centrum semiovale areas (Parkinson, Hopkins et al. 2002). Twelve percent of the patients had WM hyperintensities on T2WI on day 1 (Parkinson, Hopkins et al. 2002) with significantly more periventricular, but not centrum semiovale distributions as compared with age-matched controls. The WM lesions in the CO group did not change from day 1 to 6 months follow-up, however the hyperintensities in the centrum semiovale were related to worse cognitive performance. The study revealed no correlation between WM hyperintensities and carboxyhemoglobin level, or duration of CO exposure at any of the three scan times (Parkinson, Hopkins et al. 2002).

Hyperintensities in T2WI and fluid-attenuated inversion recovery (FLAIR) and hypointensities in T1WI often suggest WM demyelination or axonopathy (Chang, Han et al. 1992; Pavese, Napolitano et al. 1999; Parkinson, Hopkins et al. 2002). From a pathological perspective, myelin damage is constant and can vary from discrete perivascular lesions to extensive periventricular demyelination and/or axonal destruction (Funata, Okeda et al. 1982; Prockop and Chichkova 2007). An autopsy study after CO intoxication showed that diffuse WM hyperintensities reflected apoptosis of oligodendrocytes (Akaiwa, Hozumi et al. 2002). Another autopsy study of brains three days after CO intoxication revealed a normal cortex and injured WM with disrupted myelin and pyknotic oligodendroglia, whilst the axons, astrocytes and capillaries were normal (Foncin and Le Beau 1978).

Fig. 6. A wide spectrum of white matter hyperintensities in fluid-attenuated inversion recovery after carbon monoxide intoxication with cognitive deficits.

Neuroimaging Studies in Carbon Monoxide Intoxication 363

In the acute phase, petechial hemorrhages of the WM, particularly the corpus callosum, are common (Funata, Okeda et al. 1982; Finelli and DiMario 2004; Weaver and Hopkins 2005). Gradient echo T2WI uses a shorter repetition time than spin-echo T2WI and can detect metal material such as ferritin and ferritin-containing substances such as hemosiderin, thus detecting hemorrhages and microbleeds (Atlas, Grossman et al. 1988; Bradley 1993). Susceptibility-weighted imaging (SWI) is a heavy T2\*-weighted gradient-recalled 3-D fast low-angle shot sequence with full flow compensation in all three directions (Sehgal, Delproposto et al. 2005). Microhemorrhages have been reported in patients with CO intoxication with the complimentary information provided by gradient echo T2WI and SWI (Finelli and DiMario 2004; Weaver and Hopkins 2005). In gradient echo T2WI, hemorrhages along the nerve fibers are distributed predominantly over the posterior WM (Finelli and

**4.2.3 Imaging features suggesting WM hemorrhage** 

Fig. 8. Microhemorrhage shown on susceptibility-weighted imaging.

**4.3.1 Imaging features suggesting cortical injury and atrophy** 

attenuated inversion recovery (8D).

**4.3 Cortex** 

Four months after carbon monoxide intoxication, a 53-year-old woman with a low signal intensity lesion on susceptibility-weighted imaging (8A, arrow) suggesting microhemorrhage of white matter which was invisible on T1 (8B), T2 (8C), and fluid-

Pure cortical involvement without concurrent WM lesions in CO intoxication is not common (Choi, Kim et al. 1993). Using DWI, imaging features suggesting cortical cytotoxic edema were described in bilateral posterior temporal lobes and bilateral occipital lobes in one patient, bilateral posterior temporal lobes and left parietal lobe in

DiMario 2004).

Focal white matter hyperintensities (WMHs) over bilateral frontal horns in a 29-year-old woman, two years after CO exposure (6A). Diffuse and confluent WMHs in a 42-year-old woman, one and a half months after CO exposure (6B). Prominent subcortical U fiber hyperintensity with globus pallidus hyperintensity in a 35-year-old man, one and a half months after CO exposure (6C). A 31-year-old woman presented in a confused state without obvious WMHs four days after CO intoxication (6D). Extensive subcortical WMHs with globus pallidus hypointensity two years later (6E).

A study by Weaver (Weaver, Valentine et al. 2007) suggested that cognitive sequelae at six weeks benefited from hyperbaric oxygen (HBO) in patients aged 36 years and older, or who were exposed to CO for a duration of 24 hours or more. Two studies explored changes of fractional anisotropy (FA) in CO intoxication after HBO. Both studies revealed lower FA values in the patient group compared to that of controls three months after HBO (Lo, Chen et al. 2007; Chang, Lee et al. 2009). The mini-mental state examination scores completely recovered after three months of follow-up in all evaluated patients in one study (Lo, Chen et al. 2007), while another study showed that HBO treatment may not reverse the damage caused by CO intoxication (Chang, Lee et al. 2009). A longitudinal study used diffusion tensor imaging (DTI) and compared the changes of diffusion measurements in CO intoxication patients including mean diffusivity, axial diffusivity and radial diffusivity with follow-up scans three months and 10 months later. Extensive changes found in the FA maps at both three and 10 months in the CO group were attributed to initial increments of radial diffusivities, while a decrement of axial diffusivities were found at 10 months follow-up (Chang, Chang et al. 2010). The study suggested that changes in diffusion parameters might reflect WM demyelination at three months followed by subsequent axonopathy.

Fig. 7. An example of Tract Based Spatial Statistics with decreased Fractional Anisotropy (FA) (blue) overlaid on the mean FA skeleton (green) in a sample of carbon monoxide intoxication (n=30) as compared with age-matched controls. Diffuse white matter damage was detected including the subcortical areas, brain stem and cerebellum.

White matter insults after CO intoxication lead to transient or permanent injuries, which consequently lead to decreased WM volumes. Diffusion indices including mean diffusivity, axial diffusivity and radial diffusivity reflect WM injuries earlier than volume reduction, while the major regions of WM atrophy in one study were in the periventricular WM areas (Chang, Chang et al. 2010).

Focal white matter hyperintensities (WMHs) over bilateral frontal horns in a 29-year-old woman, two years after CO exposure (6A). Diffuse and confluent WMHs in a 42-year-old woman, one and a half months after CO exposure (6B). Prominent subcortical U fiber hyperintensity with globus pallidus hyperintensity in a 35-year-old man, one and a half months after CO exposure (6C). A 31-year-old woman presented in a confused state without obvious WMHs four days after CO intoxication (6D). Extensive subcortical WMHs with

A study by Weaver (Weaver, Valentine et al. 2007) suggested that cognitive sequelae at six weeks benefited from hyperbaric oxygen (HBO) in patients aged 36 years and older, or who were exposed to CO for a duration of 24 hours or more. Two studies explored changes of fractional anisotropy (FA) in CO intoxication after HBO. Both studies revealed lower FA values in the patient group compared to that of controls three months after HBO (Lo, Chen et al. 2007; Chang, Lee et al. 2009). The mini-mental state examination scores completely recovered after three months of follow-up in all evaluated patients in one study (Lo, Chen et al. 2007), while another study showed that HBO treatment may not reverse the damage caused by CO intoxication (Chang, Lee et al. 2009). A longitudinal study used diffusion tensor imaging (DTI) and compared the changes of diffusion measurements in CO intoxication patients including mean diffusivity, axial diffusivity and radial diffusivity with follow-up scans three months and 10 months later. Extensive changes found in the FA maps at both three and 10 months in the CO group were attributed to initial increments of radial diffusivities, while a decrement of axial diffusivities were found at 10 months follow-up (Chang, Chang et al. 2010). The study suggested that changes in diffusion parameters might

reflect WM demyelination at three months followed by subsequent axonopathy.

Fig. 7. An example of Tract Based Spatial Statistics with decreased Fractional Anisotropy (FA) (blue) overlaid on the mean FA skeleton (green) in a sample of carbon monoxide intoxication (n=30) as compared with age-matched controls. Diffuse white matter damage

White matter insults after CO intoxication lead to transient or permanent injuries, which consequently lead to decreased WM volumes. Diffusion indices including mean diffusivity, axial diffusivity and radial diffusivity reflect WM injuries earlier than volume reduction, while the major regions of WM atrophy in one study were in the periventricular WM areas

was detected including the subcortical areas, brain stem and cerebellum.

(Chang, Chang et al. 2010).

globus pallidus hypointensity two years later (6E).
