Книга_5813_Библиотеки_им_академика_М_И_Перельмана

Добавил:

Sekretar kiopkiopkiop18@yandex.ru t.me/Prokururor I Вовсе не секретарь, но почту проверяю Опубликованный материал нарушает ваши авторские права? Сообщите нам.

Вуз:

Ростовский Государственный Медицинский Университет

Предмет:

Медицина общая

Файл:

Ординатура / Хирургия / Библиотека им академика М.И. Перельмана / Книга_5813_Библиотеки_им_академика_М_И_Перельмана.pdf

Скачиваний:

Добавлен:

02.09.2026

Размер:

21 Мб

Скачать

☆

<<< < Предыдущая 1 2 3 4 5 6 7 8 910 / 3210 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 > Следующая >>>

Specicity

(%) AUC

nICP method accuracy and

correlation measures with ICP Sensitivity (%)

prior to shunting was

signicantly increased.

Pre-shunting ICP and PI were

not correlated (R=0.37)

with midline shift as indication

)

of elevated ICP (R=0.66

(threshold

10mmHg)

(threshold

88 (threshold of

10mmHg)

83 (threshold of

20mmHg)

(ICP=10.93×PI−1.28)

In the ICP range of

5–40mmHg, the correlation

was R=0.94

formula is:

20mmHg)

)

=0.73

In this interval, SD for nICP_PI

was ±2.5mmHg in the ICP

ICP=11.5×PI–2.23

ICP and PI was R=0.64

range of 5–40mmHg

95% CI of ±4.2mmHg

D. Cardim and C. Robba

=0.22

For ICP ≥20mmHg,

correlation was R=0.82aPI allows early identication of

patients with low CPP and risk

(ICP=23×PI+14)

of cerebral ischaemia

was R

95% CI for a mean ICP of

20mmHg was −3.8 to

43.8mmHg

PI is not a reliable predictor of

ICP

Invasive ICP

monitoring

Epidural PI in patients with elevated ICP

Sample size and

disease

29,

Rainov etal. [59] Investigate a

Method Author Study purpose

Table 8.1 (continued)

Hydrocephalus

possible

relationship

between PI, RI, FV

and ICP changes in

adult patients with

NM Increases in PI were correlated

18, Stroke and

MCA infarction

hydrocephalus.

with clinical

Asil etal. [60] TCD was compared

Intraventricular Correlation between PI and ICP

81 (SAH, TBI

and other

intracranial

relationship

ndings.

between ICP and PI

examination and

neuroradiologic

Bellner etal. [61] Investigate the

disorders)

in neurosurgical

patients.

37, TBI Intraparenchymal Overall correlation between

a tool for detection

of cerebral

haemodynamic

Voulgaris etal. [62] Investigate TCD as

10, INPH Intraparenchymal Correlation between PI and ICP

changes

Behrens etal. [63] Validate TCD as a

method for ICP

determination of

INPH

8 Transcranial Doppler andOptic Nerve Ultrasonography forNon-invasive ICP Assessment

(threshold

20mmHg)

25 (threshold of

20mmHg)

No signicant relationships

between PI and ICP when

differences within individuals

or binary examination of PI (PI

<1 and ≥1) were considered

PI and ICP of R=0.36

PI is not a reliable non-invasive

95 (for

detecting

initial ICH)

94 (for

detecting initial

ICH)

nICP similar to measured ICP,

with nICP of 10.6±4.8 and

indicator of ICP in children

with severe TBI

ICP of 10.3±2.8mmHg

PI ≥1.31 was observed in 94%

of cases with initially elevated

ICP, and 59% of those with

0.62 (ICP

≥15mmHg)

normal initial ICP values

TCD is an excellent rst-line

examination to screen children

who need urgent treatment and

continuous invasive ICP

monitoring

was R=0.31

0.74 (ICP

≥35mmHg)

95% Prediction interval

>±15mmHg

The value of PI to assess nICP

is very limited

in such situations was R=0.70

(continued)

95% CI of ±21mmHg

NM Marginal correlation between

34 Children,

TBI

relationship

between PI, ICP

and CPP in children

with severe TBI

Figaji etal. [64] Examine the

Intraparenchymal

45, TBI Intraventricular Bellner’s equation resulted in

Brandi etal. [65] Assess an optimal

117 Children,

TBI

nICP and nCPP

following TBI using

TCD.

Evaluate the

accuracy of TCD in

Tude Melo etal.

[66]

emergency settings

to predict

intracranial

hypertension and

abnormal CPP in

290, TBI Intraparenchymal Correlation between PI and ICP

diagnostic tool for

nICP and nCPP

children with TBI.

estimation.

Zweifel etal. [67] Assess PI as a

345, TBI Intraparenchymal Correlation between PI and ICP

relationship

between PI and

CVR in situations

where CVR

increases (mild

hypocapnia) and

decreases (plateau

waves of ICP) in

TBI patients.

De Riva etal. [8] Assess the

Specicity

(%) AUC

81.1 96.3 0.84

nICP method accuracy and

correlation measures with ICP Sensitivity (%)

(threshold

≥20cmH

Binomial logistic regression

indicated a strong signicant

relationship between raised ICP

and PI (OR: 2.44; 95% CI:

1.57–3.78)

(CSF pressure) was R=0.65

with rapid elevations of PI.On

Initial 24h

post-injury

Initial 24h

post-injury

100 (threshold

recording day 4, PI was

reported to be 1.93 (considering

a normal range of 0.6–1.2). ICP

was deemed elevated according

to clinical status (level of

consciousness, headache) and

papilledema

Initial 24h post-injury

Correlation between PI and ICP

was R=0.6

(threshold

≥20

mmHg for

PI of 1.3)

diastolic ow velocity, ICH intracra-

≥20mmHg for

PI of 1.3)

Beyond 24h

post-injury

47 (threshold

≥20mmHg for

PI of 1.3)

Beyond 24h post-injury

Correlation between PI and ICP

was R=0.38

D. Cardim and C. Robba

Invasive ICP

monitoring

LP Correlation between PI and ICP

Sample size and

disease

78,

Miscellaneous

intracranial

disorders

correlation between

PI with CSF

pressure.

Wakerley etal. [68] Assess the

NM Increasing ICP was associated

Case report,

sagittal sinus

thrombosis

where TCD serves

as an effective tool

for nICP

monitoring.

Wakerley etal. [69] Present a case

Intraventricular/

intraparenchymal

36 Children,

TBI

and ICP in children

with severe TBI.

relationship

between PI, FV

O’Brien etal. [70] Determine the

coefcient of determination, RI resistance index, SAH subarachnoid haemorrhage, SD

jugular bulb venous blood oxygen saturation, TBI traumatic brain injury

Method Author Study purpose

Table 8.1 (continued)

ABP arterial blood pressure, AUC area under the curve, CI condence interval; CT computerised tomography, CSF cerebrospinal uid, FV

Correlation coefcient is signicant at the 0.05 level

nial hypertension, INPH idiopathic normal pressure hydrocephalus, LP lumbar puncture, MCA middle cerebral artery, NA not available, NM not measured, NPV negative predic-

tive value, OR odds ratio, PPV positive predictive value, R correlation coefcient, R

standard deviation, SJO

()

8 Transcranial Doppler andOptic Nerve Ultrasonography forNon-invasive ICP Assessment

8.3 Methods Based ontheCalculation ofNoninvasive CPP (nCPP)

Many authors have also investigated methods based on the primarily intended calculation of noninvasive cerebral perfusion pressure (nCPP), and secondarily calculating non-invasive ICP based on the assumption that “nICP=ABP−nCPP”. nICP methods based on TCD-derived cerebral perfusion pressure are listed in Table8.2.

8.3.1 Aaslid etal. [12] (nICP

Aaslid etal. (1986) [12] were the pioneers in the development of a mathematical model for non- invasive estimation of CPP based on TCD waveform analysis. Supported by the knowledge that increased ICP was related to decreases of the ow waveform in the internal carotid artery [13], the authors deemed feasible that less evident changes in the ow pattern could be detected by more rened methods of waveform analysis. TCD, in such a perspective, could be used to obtain an estimate of ICP. Considering CPP as the driving force to ow through the cerebral vascular bed and the main factor determining pulsatile dynamics, the ow would be approximately proportional to CPP [12].

During increases in ICP, the proportion of the systolic to the diastolic CPP increases as ICP approaches the diastolic ABP.Consequently, the pulsatile component of CPP increases proportionally to its mean value and this would also be reected in the FV waveform. Thus, the ratio between mean FV (FVm) and pulsatile amplitude of FV (f1) would be expected to be related to CPP.The simplest approach, in this case, would be using this ratio as an index for CPP.However, it does not consider changes in the ratio caused by variations in the amplitude of ABP waveform. Such limitation was solved by multiplying the ratio described above by the amplitude of the rst harmonic of ABP (a1) [12]:

nCPP FV mmHg

Aaslid m

´/ fa11

Aaslid

)

This method was tested in patients with supratentorial hydrocephalus undergoing ventricular infusion tests. The linear regression determined for nCPP estimation considering all cases was

CPPnCP

11 5.

Aaslid

mHg

In the individual case, there was a strong correlation between nCPP and CPP, R=0.93–0.99. The standard deviation between nCPP and CPP was

8.2mmHg at 40mmHg, while the mean deviation was only 1mmHg. Overestimation occurred in two cases, 10 mmHg and 18 mmHg, respectively. Underestimation was present in other two cases, 9 and 8mmHg, respectively. For the six remaining cases, the estimates were within ±5mmHg of measured CPP. The method could differentiate correctly between low (<40 mmHg) and normal (>80mmHg) CPP in all patients. At higher levels of CPP, the estimates presented with low accuracy, showing correct differentiation between CPPs of 70 and 100mmHg in 80% of the cases.

8.3.2 Czosnyka etal. [14] (nICP

FVd

)

Some studies have demonstrated that certain patterns of the TCD waveform, like a decrease in FVd, reect impaired cerebral perfusion caused by a CPP decrease [14, 15]. Czosnyka et al. [14] described this relationship as the following formula:

nCPP ABP

=´+ 14mmHg

The correlation between nCPP and invasively measured CPP in traumatic brain injury (TBI) patients was R= 0.73, p<0.001. In 71% of the examinations, the estimation error was below 10mmHg, and in 84% of the examinations, the error was less than 15mmHg. The method had a high positive predictive value (94%) for detecting low CPP (<60 mmHg). Estimated CPP was highly specic for detecting changes in measured CPP over time, caused by either increases in ICP (plateau waves) or systemic hypotension. In six patients presenting plateau waves, nCPP compared with measured CPP had an average goodness- of-t coefcient R2 of 0.82.

94 (CPP

≤60mmHg)

D. Cardim and C. Robba

(≥20mmHg)

nICP/nCPP method accuracy and

correlation measures with ICP/CPP AUC PPV (%) NPV (%)

40mmHg

Able to differentiate between low

(≤40mmHg) and normal (≥80mmHg) CPP

in all cases

Low accuracy (80%) at higher levels of CPP

Invasive ICP

monitoring

Intraventricular SD for nCPP estimation of 8.2mmHg at

Sample size and

disease

10, Supratentorial

hydrocephalus

Sensitive to detect changes of CPP over time

(70–100mmHg)

. 96, TBI Intraparenchymal 95% PE for nCPP estimation was >27mmHg

Aaslid

Describe and assess a

method for nCPP

calculation based on

FV and ABP.

Aaslid etal.

[12]

Assess nICP

Czosnyka

etal. [14]

Estimation error was less than 10 and

R=0.73

=0.82

15mmHg in 71% and 84% of the cases,

respectively

Highly specic for detecting changes over

time: R

Averaged correlation considering day-by-day

variability between CPP and nCPP was

R=0.71

96, TBI Intraparenchymal Correlation between nCPP and CPP was

Describe and assess a

method for nCPP

calculation based on

FV (using the

concept of FVd) and

ABP.

Czosnyka

etal. [14]

Mean values of nCPP and CPP were

66.10±10.55mmHg and

and 13mmHg in 89% and 92% of the cases,

respectively

95% CI of ±12mmHg

R=0.92

65.40±10.03mmHg, respectively

Bias of 5.6mmHg and 95% CI of

±17.4mmHg

nCPP:

45, TBI Intraparenchymal nICP:

. 25, TBI Intraparenchymal Error for nCPP estimation was less than 10

FVd

Assess nICP

Schmidt

. 47, TBI Intraparenchymal Correlation between nCPP and CPP was

FVd

Assess an optimal

nICP and nCPP

following TBI using

Assess nICP

etal. [71]

Gura etal.

[72]

TCD.

Brandi etal.

[65]

Bias of 6.2mmHg and 95% CI of 11.8mmHg 0.96

Bias of −5.5mmHg and 95% CI of

±20.6mmHg

Intraparenchymal,

10 intraventricular

injury

. 38, Acute brain

FVd

Assess nICP

Rasulo etal.

[73]

Aaslid

Method Author Study purpose

Table 8.2 nICP methods based on non-invasive cerebral perfusion pressure estimation

nICP

FVd

nICP

8 Transcranial Doppler andOptic Nerve Ultrasonography forNon-invasive ICP Assessment

0.91

(≥20mmHg)

0.34

(≥20mmHg)

(R=−0.17)

R=0.30

nCPP and CPP were correlated (slope, 0.76;

intercept, 10.9; 95% CI, −3.5 to 25.4mmHg)

During hypercapnia:

nCPP and CPP were correlated, but with

increased discrepancy, as reected in

Intraparenchymal Correlation between ICP and nICP was

ischaemic brain

injury after

cardiac arrest

20, TBI Intraparenchymal During normocapnia:

. 100, TBI Intraparenchymal No correlation between ICP and nICP

. 11, Hypoxic-

FVd

reactivity

>0.8

Bias of 6.8mmHg and 95% CI ±19.7mmHg

nCPP:

condence interval (slope, 0.55; intercept,

32.6; 95% CI, 16.3–48.9mmHg)

Bias of −6.8mmHg and 95% CI ±45.2mmHg

Bias ± SD of 4.02±6.01mmHg

R=0.85

280, TBI Intraparenchymal Correlation between nCPP and CPP was

. 45, TBI Intraparenchymal nICP:

Edouard

nCPP estimation error was below 10mmHg in

83.3% of the cases

Temporal analysis:

Mean correlation in time domain was R=0.73

(0.23–0.99)

Correlation between nCPP and CPP was

Bias of 3.45mmHg (range: 4.69–9.03mmHg)

Mean SD of 5.52mmHg (range: 1.52–

R=0.67

10.76mmHg) and 95% CI of the SD of

1.89–5.01mmHg

Bias of 19.61mmHg

95% CI: ± 40.1mmHg

Temporal analysis:

Mean correlation in time domain was R=0.55

(±0.42)

coefcient of determination, SD standard deviation, SAH subarachnoid haemorrhage, TBI traumatic brain injury

Assess nICP

Cardim etal.

[74]

Cardim etal.

[42]

Describe and assess a

method for nCPP

calculation based on

Edouard

etal. [16]

Edouard

test.

FV and ABP under

stable conditions and

during CO

nICP

Assess nICP

Brandi etal.

[65]

Describe and assess a

method for nCPP

calculation based on

FV (using the

concept of CrCP) and

ABP.

Varsos etal.

[18]

CrCP

nICP

Describe and assess a

method for nCPP

calculation based on

FV and ABP.

Abecasis &

Cardim etal.

[24]

Spectral

nCPP

AUC area under the curve, ABP arterial blood pressure, CI condence interval, FV cerebral blood ow velocity, NPV negative predictive value, PPV positive predictive value,

PE prediction error, R correlation coefcient, R

Correlation coefcient is signicant at the 0.05 level

()

D. Cardim and C. Robba

Furthermore, nCPP was very sensitive to detecting decreases in ABP below 70 mmHg (in ten patients) with an average R2 of 0.92. A good correlation was found between the average measured CPP and nCPP when day-by-day variability was assessed in a group of 41 patients (R=0.71).

8.3.3 Edouard etal. [16] (nICP

Edouard

)

This method is based on a non-invasive assessment of CPP using a combination of phasic values of both FV in the MCA and

nCPP

Edouard

FV FV

ABPm and ABPd represent the mean and dia-

stolic ABP, respectively.

Twenty adults with bilateral and diffuse brain injuries were included in the study and subdivided into two groups. In group A (N=10), the comparison was repeatedly performed under stable conditions. In group B (N=10), the comparison was conducted during a CO2 reactivity test. nICP was not estimated in this study. In group A, nCPP and measured CPP were correlated (slope,

0.76; intercept, +10.90; 95% CI, −3.50 to +25.40). The relationship persisted during ICP increase caused by the reactivity test in group B (slope, 0.55; intercept, +32.60; 95% CI, +16.30 to +48.90). However, the discrepancy between nCPP and measured CPP increased as reected by the increase in bias and variability.

8.3.4 Varsos etal. [18] (nICP

CrCP

)

The concept of critical closing pressure (CrCP) was first introduced by Burton’s model,

ABP.It was primarily validated in preeclamptic and healthy pregnant women by comparing nCPP with CPP measured at the epidural space [17].

In Edouard etal.’s prospective study [16], the objective was to assess the adequacy of such method for TBI patients with invasive ICP monitoring, in both a stable state and during a rapid change in cerebrovascular tone following an induced alteration in arterial blood carbon dioxide pressure (PaCO2).

The non-invasive CPP (nICP

Edouard

) was calcu-

lated using the following formula:

ABPABP mmHg

´-

÷ ø

described as the sum of ICP and vascular wall tension (WT) [19]. WT represents the active cerebral vasomotor tone that combined with ICP determines the CrCP. Clinically, CrCP represents the lower threshold of ABP below which blood pressure in the brain microvasculature is inadequate to prevent the collapse and cessation of blood flow [19]. CrCP can be assessed non-invasively using TCD, by comparing the pulsatile waveforms of FV and ABP.Given the relationship with the vasomotor tone of small blood vessels, CrCP can provide information regarding the state of cerebral haemodynamics and reflect changes in CPP [19–23].

The conception and assessment of the model for the non-invasive estimator of CPP (nCPP

) were performed using a cohort of

CrCP

280 TBI patients, divided into two subgroups: the formation group with 232 patients (including 455 recordings) and the validation group with 48 patients (including 325 recordings):

nCPP ABP

=´ -

CrCP

é ê

ê ê

()

0 266

CVRC HR

×× ×

ù ú

-0 734

7 026

mmHg

ABP

CBV

8 Transcranial Doppler andOptic Nerve Ultrasonography forNon-invasive ICP Assessment

CVR

CVR (mmHg/(cm/s)) represents cerebral vas-

cular resistance, C

(cm/mmHg) denotes compli-

ance of the cerebral arterial bed (arteries and arterioles) and HR represents heart rate given in beat/s. a1 represents the pulse amplitude of the rst harmonic of the ABP waveform; f1 the pulse amplitude of the rst harmonic of the FV waveform; and CaBV1 the pulse amplitude of the rst harmonic of the cerebral arterial blood volume waveform (CaBV). The pulse amplitude of rst harmonics is determined with fast Fourier transformation (FFT).

nCPP

using data from the validation group. nCPP

was tested against invasive CPP

CrCP

was correlated with measured CPP (R = 0.85, p<0.001), with a bias of 4.02±6.01mmHg, and in 83.3% of the cases with an estimation error below 10mmHg. nCPP

prediction analysis at

CrCP

low CPP limits (50, 60 and 70mmHg) resulted in AUCs greater than 0.8 for all limits. nCPP

CrCP

was found to be strongly correlated with CPP changes in the time domain (mean R = 0.73, range 0.23–0.99). For each patient, nCPP

CrCP

presented a mean difference from CPP of 3.45mmHg (range 4.69–19.03mmHg), and a mean standard deviation of this difference of 5.52mmHg (range

1.52–10.76mmHg). nCPP

could predict CPP

CrCP

between multiple recording sessions with a 95% CI of 1.89–5.01mmHg.

8.3.5 Abecasis andCardim etal.

(Spectral nCPP) [24]

This recently proposed method is based on accounting for changes in pulsatile cerebral arterial blood volume (CaBV) applied to the analysis of a simplied hydrodynamic model of CBF and CSF dynamics [25–27]. The spectral feature of the method has some advantages: it creates partial independence from the inaccuracy associated with zeroing ABP transducers at heart level (a common clinical practice) and partially elimi-

nates the issue of the time delay between peripheral ABP and FV in the MCA.The dynamical model of CSF and cerebral blood circulation proposed by Ursino and Lodi [25, 26] is the theoretical basis for the spectral nCPP method; the former has also been modied by other authors to incorporate changes in cerebral blood volume and adapted to nICP or nCPP estimations. nICP

[28] described in Sect. 8.4 is one exam-

Heldt

ple. Similarly to this method, spectral nCPP produces patient-specic CPP estimates and does not require calibration datasets of reference patients.

Abecasis and Cardim etal. [24] assessed spectral nCPP on a retrospective analysis from 19 children (69 TCD recordings). Spectral nCPP presented good correlation with CPP measure invasively (R= 0.67 (p< 0.0001)), and a good mean correlation in time domain (R=0.55±0.42). Spectral nCPP also demonstrated strong ability to predict values of CPP below 70mmHg (AUC of 0.908 (95% CI=0.83–0.98)). The agreement between spectral nCPP and invasive CPP assessed by Bland-Altman analysis revealed that nCPP overestimated invasive CPP by 19.61 mmHg with a wide 95% CI of ±40.4mmHg.

8.4 Methods Based

onMathematical Models Associating Cerebral Blood Flow Velocity andArterial Blood Pressure

Several authors have proposed mathematical models that simulate the cerebrovascular dynamics using simultaneous FV and ABP measurements. nICP methods based on mathematical models are listed in Table8.3.

8.4.1 Black-Box Model forICP

Estimation (nICP

In this model, the intracranial compartment is considered a black-box (BB) system, with ICP being a system response to the incoming signal ABP [29]. This mathematical model originates

)

Specicity

(%) AUC

Sensitivity

(%)

CSF

D. Cardim and C. Robba

For Mx: 92

For PRx: 67

For Mx: 97

CSF

For PRx: 61

and R

CSF

nICP method accuracy and

correlation measures with ICP

1.8mmHg

In this cohort a maximum 95% CI of

±12.8mmHg was found

Correlation between nR

was R=0.73

Invasive ICP

monitoring

Sample size and

disease

11, TBI Epidural Bias of 4.0mmHg and SDE of

Epidural Analysis across all records:

21, Different

types of

hydrocephalus

and R

CSF

prediction of 2.2mmHg minute/ml

Analysis specic to different

subtypes of hydrocephalus:

Bias of 4.1mmHg and SDE for R

Correlation between nR

was R=0.89

during ICP increase was R=0.98

prediction of 1.7mmHg min/ml

Bias of 2.7mmHg and SDE for R

Intraparenchymal Correlation between nICP and ICP

17, TBI (plateau

(A) waves

during plateau

Analysis considering the baseline of

plateau waves:

Bias of 8.3mmHg and SDE of

5.4mmHg

Analysis considering the top of

plateau waves:

Bias of 7.9mmHg and SDE of

4.3mmHg

Increases of nICP and direct ICP

Correlation between nMx and Mx

during plateau waves:

R=0.98 (p<0.001)

Intraparenchymal/

observed in 7

patients)

145 (135 TBI, 10

was R=0.90

Correlation between nPRx and PRx

was R=0.62aMedian bias of 6.0mmHg

Only TBI:

Bias of 7.1mmHg

Only stroke:

Bias of 4.3mmHg

intraventricular

stroke)

using nICP

Describe and assess a method

for nICP calculation based on

FV and ABP using a black-box

model.

This study aimed at predicting

Schmidt

etal. [29]

Schmidt

Method Author Study purpose

Table 8.3 nICP methods based on mathematical models

nICP

CSF

the time course of raised ICP

during CSF infusion tests and

its suitability for estimating the

etal. [75]

waves of ICP.

Assess nICP

Schmidt

etal. [76]

to adapt to the SCA,

using Mx and PRx as

This study aimed at

Schmidt

parameters

investigating the ability of

nICP

etal. [77]

8 Transcranial Doppler andOptic Nerve Ultrasonography forNon-invasive ICP Assessment

0.83

(≥20mmHg)

0.88

(≥20mmHg)

(linear)

0.92

resistance to cerebrospinal

CSF

Correlation between nICP and ICP

was R=0.90

Bias of 1.6 and SDE of ±7.6mmHg

Inferred 95% CI of ±14.9mmHg

Across bilateral TCD records:

Correlation between nICP and ICP

was R=0.76

Bias 1.5 and SDE 5.9mmHg

Inferred 95% CI of ±11.6mmHg

On a patient record basis:

Intraparenchymal Across all TCD records:

37 (45 TCD

recordings in

total, 30 bilateral),

TBI

Describe and assess a method

for nICP calculation based on

FV and ABP.

Kashif and

Heldt etal.

[28]

Biases for non-linear models were

<6.0mmHg compared to 6.7mmHg

Correlation between nICP and ICP

was R=0.90 (N=45 recordings)

of the nICP

Inferred 95% CIs for non-linear

models were ≤±10.8mmHg,

Intraparenchymal/

intraventricular

23 (14 TBI, 9

hydrocephalus)

Describe and assess a method

for nICP calculation based on

FV and ABP, using an

appropriated model for nICP

Xu etal.

[33]

mining nICP and ICP was R=0.80

compared to 10.6mmHg of the

linear model

9, TBI Intraventricular Median correlation between data

Describe and assess a method

for nICP calculation based on

[38]

Decision curve analysis showed that

method presented a median bias of

57, TBI NM Kernel spectral regression-based

FV and ABP based on the

concepts of data mining.

Describe and assess a method

for nICP calculation based on

Kim etal.

[39]

the semi-supervised method is more

4.37mmHg

Intraparenchymal/

intraventricular

90 (44 TBI, 36

SAH, 10 NPH)

Describe and assess a method

for nICP calculation based on

FV and ABP based on the

concepts of data mining.

Kim etal.

[40]

accurate and clinically useful than

the supervised or PI-based method

coefcient of determination, SDE standard deviation of the error, R

; NPH normal pressure hydrocephalus, TBI traumatic brain injury, SAH subarachnoid haemorrhage, SCA state of cerebral

CSF

FV and ABP based on the

concepts of semi-supervised

machine learning.

nICP-derived R

CSF

Heldt

nICP

Modied

nICP

Data mining Hu etal.

Semi-

supervised

learning

Correlation coefcient is signicant at the 0.05 level

Inferred 95% CI was calculated as 1.96*SDE

AUC area under the curve, CI condence interval, R correlation coefcient, R

uid (CSF) outow, R

autoregulation

<<< < Предыдущая 1 2 3 4 5 6 7 8 910 / 3210 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 > Следующая >>>

Соседние файлы в папке Библиотека им академика М.И. Перельмана

#
31.08.202624 Мб0Книга_5809_Библиотеки_им_академика_М_И_Перельмана.pdf
#
30.08.202677 Мб0Книга_581_Библиотеки_им_академика_М_И_Перельмана.pdf
#
31.08.202624 Мб0Книга_5810_Библиотеки_им_академика_М_И_Перельмана.pdf
#
31.08.202623 Мб0Книга_5811_Библиотеки_им_академика_М_И_Перельмана.pdf
#
31.08.202622 Мб0Книга_5812_Библиотеки_им_академика_М_И_Перельмана.pdf
#
02.09.202621 Мб0Книга_5813_Библиотеки_им_академика_М_И_Перельмана.pdf
#
02.09.202621 Мб0Книга_5814_Библиотеки_им_академика_М_И_Перельмана.pdf
#
02.09.202620 Мб0Книга_5815_Библиотеки_им_академика_М_И_Перельмана.pdf
#
02.09.202619 Мб0Книга_5816_Библиотеки_им_академика_М_И_Перельмана.pdf
#
30.08.202677 Мб0Книга_582_Библиотеки_им_академика_М_И_Перельмана.epub
#
30.08.202677 Мб0Книга_583_Библиотеки_им_академика_М_И_Перельмана.pdf