We evaluated whether an irrigated contact force–sensing catheter would improve the safety and effectiveness of radiofrequency ablation of premature ventricular contractions originating from the right ventricular outflow tract.

We retrospectively reviewed the charts of patients with symptomatic premature ventricular contractions who underwent ablation with a contact force–sensing catheter (56 patients, SmartTouch) or conventional catheter (59 patients, ThermoCool) at our hospital from August 2013 through December 2015. During a mean follow-up of 16 ± 5 months, 3 patients in the conventional group had recurrences, compared with none in the contact force group. Complications occurred only in the conventional group (one steam pop; 2 ablations suspended because of significantly increasing impedance). In the contact force group, the median contact force during ablation was 10 g (interquartile range, 7–14 g). Times for overall procedure (36.9 ± 5 min), fluoroscopy (86.3 ± 22.7 s), and ablation (60.3 ± 21.4 s) were significantly shorter in the contact force group than in the conventional group (46.2 ± 6.2 min, 107.7 ± 30 s, and 88.7 ± 32.3 s, respectively; P <0.001). In the contact force group, cases with a force-time integral <560 gram-seconds (g-s) had significantly longer procedure and fluoroscopy times (both P <0.001) than did those with a force-time integral ≥560 g-s.

These findings suggest that ablation of premature ventricular contractions originating from the right ventricular outflow tract with an irrigated contact force–sensing catheter instead of a conventional catheter shortens overall procedure, fluoroscopy, and ablation times without increasing risk of recurrence or complications.

The creation of durable transmural lesions during radiofrequency catheter ablation (RFCA) of cardiac arrhythmias depends on several factors, including radiofrequency (RF) power, duration of energy application, electrode temperature, tip orientation, and tip size.1,2  Lesion formation also depends greatly on the contact force (CF) between catheter tip and target tissue.3–5  Insufficient CF may necessitate increased power and longer ablation times to create therapeutically effective lesions5 ; excessive CF may increase the risk of complications such as steam pop, heart wall perforation, and cardiac tamponade.6  Therefore, in theory, more precise control of CF during ablation should help produce lesions that are more reliable and effective and less prone to complications.7 

A novel open-irrigated RFCA catheter tipped with a sensor to enable real-time measurement of catheter tip-to-tissue CF has been developed3–5  and is used clinically. The multicenter TOuCh+ for CATheter Ablation (Toccata) trial8  in patients with supraventricular tachycardia and atrial fibrillation (AF) showed that ablation was safer with the CF-sensing catheter than with a conventional catheter. Other studies in patients with AF showed that ablation with the CF-sensing catheter improved lesion size and depth and thus improved outcomes.9,10  However, few studies have investigated the safety and effectiveness of the CF-sensing catheter when used for ablation of premature ventricular contractions (PVCs).9,11  Therefore, we evaluated whether an irrigated CF-sensing catheter would improve the safety and effectiveness of RFCA of idiopathic PVCs originating from the right ventricular outflow tract (RVOT).

Patients and Methods

We retrospectively reviewed the records of patients at the First Affiliated Hospital of Dalian Medical University who underwent RFCA of PVCs originating from the RVOT from August 2013 through December 2015. Patients were characterized in terms of history; results of physical examination, conventional 12-lead electrocardiography (ECG), 24-hour 12-lead Holter monitoring, exercise stress test, transthoracic echocardiography, and chest radiography; and results of electrolyte, thyroid, hepatic, and renal function laboratory tests. Inclusion criteria included the following: frequent symptomatic PVCs originating from the RVOT documented by 12-lead ECG to have an inferior axis and a left bundle branch block QRS morphology; PVC burden >25%; and first-time ablation for PVC. Excluded were patients who had a non-RVOT origin for PVCs as indicated by an S wave in lead I, an R-wave duration index ≥0.5 in leads V1 and V2, and an R/S wave amplitude index ≥0.3 in leads V1 and V2; evidence of structural heart disease, including ischemic or valvular heart disease, hypertrophy and dilated cardiomyopathy, sarcoidosis, amyloidosis, arrhythmogenic right ventricular cardiomyopathy, and congenital defects; hyperthyroidism or electrolyte disturbances; drug toxicity; diabetes mellitus; blood pressure >160/100 mmHg; impaired renal function; a QT interval >450 ms without bundle branch block; atrioventricular block, bundle branch block, or both; a history of syncope; and decreased blood pressure at the onset of ventricular tachycardia (VT). The Ethics Committee at the First Affiliated Hospital approved the study protocol.

Our retrospective review included patients who had undergone RFCA with either a CF-sensing catheter (CF group) or a conventional catheter (conventional group) at our hospital. Three operators (Drs. Xia, Yin, and Gao) had experience using both catheters and conducted the ablations independently. The CF-sensing catheter used for RFCA was a 3.5-mm Carto® ThermoCool SmartTouch® SF irrigated RF ablation catheter (Biosense Webster, a Johnson & Johnson company) equipped with a force sensor. The conventional catheter used for RFCA was a 3.5-mm ThermoCool® irrigated RF ablation catheter (Biosense Webster).

Radiofrequency Catheter Ablation

In each case, the PVC origin site, defined as the earliest site of local ventricular activation preceding onset of the QRS wave by at least 25 ms on the surface ECG, was localized as the ablation target point using the Carto 3 cardiac mapping system (Biosense Webster). During ablation, the maximum RF power setting was 40 W, and irrigated flow ranged between 2 and 17 mL/min. Termination of PVCs was confirmed by the isoproterenol infusion test.

Acute success was defined as the absence of PVCs morphologically similar to the original PVC during the first 30 min after RFCA. Recurrence was defined as the return of PVC-related symptoms and an electrocardiographic morphology similar to that seen before ablation as detected by ECG or 24-hour Holter monitoring, with a PVC burden >10% per day. For ablations done with the CF-sensing catheter, the Carto 3 mapping system was used to display the CF and 3-dimensional force vectors at 500-ms time intervals (Fig. 1). The CF value was zeroed at the inferior vena cava before mapping began. Mapping data were collected during the ablation procedure and analyzed offline to calculate the maximum, minimum, and mean CF values. The force-time integral (FTI), expressed in gram-seconds (g-s), was defined as the integral of the CF-time curve during the ablation period and was analyzed for its correlation with CF and procedure times.

Fig. 1

Shown are representative Carto 3 images captured using A) contact force (CF)–sensing SmartTouch and B) conventional ThermoCool ablation catheters in posteroanterior (left panel) and left lateral (center panel) views, along with electrical signal maps recorded by each catheter's distal and proximal electrode pairs (right panel). In A), the CF-sensing SmartTouch map in the left panel displays the target point of premature ventricular contractions (PVCs), CF value, and catheter direction (arrow). The target point is the earliest site of local ventricular activation preceding onset of the QRS wave by at least 25 ms on surface electrocardiogram. The map's display of CF value, catheter direction, and force-time integral enables the operator to know the exact position of the ablation catheter and predict lesion effectiveness. In B), the conventional ThermoCool map in the left panel displays the PVC target point; however, catheter direction must be inferred from the position of the coronary sinus catheter and other anatomic markers and, unless impedance is monitored, no information on any other parameter is available to predict lesion effectiveness.

Fig. 1

Shown are representative Carto 3 images captured using A) contact force (CF)–sensing SmartTouch and B) conventional ThermoCool ablation catheters in posteroanterior (left panel) and left lateral (center panel) views, along with electrical signal maps recorded by each catheter's distal and proximal electrode pairs (right panel). In A), the CF-sensing SmartTouch map in the left panel displays the target point of premature ventricular contractions (PVCs), CF value, and catheter direction (arrow). The target point is the earliest site of local ventricular activation preceding onset of the QRS wave by at least 25 ms on surface electrocardiogram. The map's display of CF value, catheter direction, and force-time integral enables the operator to know the exact position of the ablation catheter and predict lesion effectiveness. In B), the conventional ThermoCool map in the left panel displays the PVC target point; however, catheter direction must be inferred from the position of the coronary sinus catheter and other anatomic markers and, unless impedance is monitored, no information on any other parameter is available to predict lesion effectiveness.

Patient Follow-Up

All patients had a follow-up visit or telephone interview at 3, 6, 12, and 24 months after the initial ablation procedure. At each follow-up, the patient's history was updated, and 12-lead surface ECG and 24-hour Holter monitoring data were collected as applicable. At 24-month follow-up, patients were administered a standard questionnaire about symptoms and adverse events, including cardiovascular diseases, stroke, thrombosis, bleeding, hospitalization, and arrhythmia.

Statistical Analysis

Continuous variables were analyzed and presented as mean ± SD or as median and interquartile range and compared between the 2 study groups by using the Student t test. Categorical variables were expressed as number and percentage and were compared between the 2 groups by using the Fisher exact or χ2 test. All tests were 2-tailed. P values <0.05 were considered statistically significant. All statistical analyses were performed with SPSS, version 13.0 (SPSS, an IBM company).

Results

From August 2013 through December 2015, a total of 115 patients underwent RFCA with either a CF-sensing catheter (56 patients) or a conventional catheter (59 patients). Table I shows their clinical characteristics at baseline. Both groups were comparable in terms of age; sex; hypertension history; left ventricular ejection fraction; left atrial diameter; and levels of B-type natriuretic peptide, creatinine, thyroid-stimulating hormone, and cardiac troponin I. Overall, the mean PVC frequency was 25,316 beats; the mean PVC burden, 26.4%; and the percentage of sustained VTs (>100 beats/min for at least 30 s), 7.8%. Most PVCs (82%) originated from the septal RVOT.

Table I.

Clinical Characteristics of the Study Population (N=115)

Clinical Characteristics of the Study Population (N=115)
Clinical Characteristics of the Study Population (N=115)

Contact Force Values during Ablation

In the CF group, the median CF value during ablation was 10 g (interquartile range [IQR], 7–14 g). Moreover, the median CF values recorded at the free wall (8.6 ± 2.7 g) and septal wall (11.1 ± 4.3 g) were comparable (P=0.109). The CF and conventional groups were similar in terms of impedance (135.7 ± 18.6 vs 136.7 ± 18.3 Ω P=0.772), temperature (39.07 ± 2.9 vs 40.32 ± 4.45 °C; P=0.079), and RF power (32.9 ± 8.3 vs 30.9 ± 8.4 W; P=0.202).

Ablation Outcomes and Complications

Acute success was achieved in all patients in both study groups. No severe complications, including acute myocardial infarction, stroke, or significant bleeding, were recorded. No patient in either group had cardiac tamponade, embolization, pneumothorax, ventricular fibrillation, or arteriovenous fistula. Groin hematoma occurred in 2 patients in each group. During ablation, VT occurred in 3 patients in the CF group and 2 patients in the conventional group. All VTs were unsustained and needed no ablation or other intervention. Table II summarizes and compares the frequency of complications associated with the ablation procedure in the 2 groups.

Table II.

Complications Associated with Ablation Procedure

Complications Associated with Ablation Procedure
Complications Associated with Ablation Procedure

Contact Force Value and Ablation Safety

The importance of CF in creating reliable lesions and improving safety during AF ablation is well documented.7,10,12–15  In the CF group, VT was induced during the procedure in 3 patients (CF value >20 g). In contrast, in the conventional group (in which impedance and temperature—but not CF—were monitored), the ablation procedure had to be suspended because of increasing impedance in 2 patients. In one of these patients, the procedure was suspended because impedance rapidly increased from 120 to 200 Ω after a 10-s energy discharge. In another patient, the procedure was interrupted 36 s after the start of RF energy delivery because of a steam pop.

Contact Force and Procedure Times

Overall procedure time (36.9 ± 5 min), fluoroscopy time (86.3 ± 22.7 s), and ablation time (60.3 ± 21.4 s) were significantly shorter in the CF group than in the conventional group (46.2 ± 6.2 min, 107.7 ± 30 s, and 88.7 ± 32.3 s, respectively; all P <0.001) (Fig. 2). Post hoc multivariate regression analysis adjusting for the effects of sex, age, PVC location, and operator showed that none of these variables had any significant effect on procedure times (Table III).

Fig. 2

Graphs show that A) ablation, B) fluoroscopy, and C) procedure times were significantly shorter when a contact force–sensing catheter was used to ablate premature ventricular contractions (all P <0.001 vs conventional catheter).

Data are presented as mean ± SD. P <0.05 was considered statistically significant.

Fig. 2

Graphs show that A) ablation, B) fluoroscopy, and C) procedure times were significantly shorter when a contact force–sensing catheter was used to ablate premature ventricular contractions (all P <0.001 vs conventional catheter).

Data are presented as mean ± SD. P <0.05 was considered statistically significant.

Table III.

Adjusted Comparison of Procedure Time with and without Contact Force Sensing

Adjusted Comparison of Procedure Time with and without Contact Force Sensing
Adjusted Comparison of Procedure Time with and without Contact Force Sensing

Correlation of Force-Time Integral with Contact Force and Procedure Time

In the CF group, FTI values ranged widely from 300 to 900 g-s with a median value of 560 g-s (Fig. 3A). Linear correlation analysis indicated that CF correlated well with FTI (r=0.388, P=0.003) (Fig. 3B). With the median FTI used as cutoff, correlation analysis also indicated that the cases with a lower FTI (<560 g-s) had significantly longer procedure times (Fig. 3C) and fluoroscopy times (Fig. 3D) than did those with a higher FTI (≥560 g-s) (both P <0.001).

Fig. 3

A) Histogram shows the standard distribution of force-time integral (FTI) in the contact force (CF) group. Scatter plots show that FTI correlated linearly with B) the mean CF value, C) procedure time, and D) fluoroscopy time.

Fig. 3

A) Histogram shows the standard distribution of force-time integral (FTI) in the contact force (CF) group. Scatter plots show that FTI correlated linearly with B) the mean CF value, C) procedure time, and D) fluoroscopy time.

Patient Follow-Up

All patients were followed for 24 months. During a mean follow-up of 16 ± 5 months, 3 in the conventional group had recurrences, compared with none in the CF group. No patient reported procedure-related adverse events during telephone interviews or clinic visits.

Discussion

The results of our study suggest that using a CF-sensing catheter instead of a conventional catheter reduces ablation time during RFCA therapy for PVCs of the RVOT. They also suggest that FTI, which is known to predict RF-induced lesion size in AF ablation, correlates with procedure and fluoroscopy times during PVC ablation.

Our study supplements previous investigations into the benefits of CF sensing and measurement during cardiac ablation. Studies in bovine and porcine ex vivo models of arrhythmia have suggested that CF measurement can help predict lesion size and prevent thrombus formation, steam pops, and ablation-related complications.4,5  Emerging studies in patients with AF, supraventricular tachycardia, and ventricular arrhythmia have suggested that measuring CF during catheter ablation may reduce procedure time and risk of complications and enable more effective lesion creation.6,9,11–15  The multicenter Toccata study evaluated the real-time measurement of CF during RFCA of supraventricular tachycardia and AF8,16  and found that high CFs may occur even during catheter manipulation.8 

In agreement with a recent meta-analysis of CF-guided AF ablation,10  our study suggests that CF-guided RFCA of PVCs is as safe and effective as non-CF-guided ablation. Also, as already known from studies with CF-sensing catheters, lesion size is markedly affected by catheter contact4,5 ; the CF between catheter tip and tissue correlates with procedure duration and ablation outcome3,4 ; and impedance drop during ablation, an established indicator of complete ablation17–19  and surrogate marker of lesion formation,20  closely correlates with CF and FTI.20–22  Thus, monitoring CF during ablation would safely provide experienced operators with feedback for delivering RF energy more effectively and creating lesions more completely.

In addition, CF measurement could be useful for training purposes. Operators could learn to use the immediate feedback on catheter manipulation to reduce variability in CF application. For instance, because the catheter forms an angled curve inside the RVOT to accommodate the RVOT's peculiar anatomy, the operator could use measurements by a CF-sensing catheter to gauge whether tip-to-tissue contact is adequate and whether therapeutically effective lesions are being created. High CF values may occur even while the catheter is being manipulated for mapping or ablation, so CF monitoring could help to avoid complications during the procedure.6 

Study Limitations

Our study had several important limitations. First, it only included patients with PVCs originating from the RVOT, thus limiting any conclusions to that particular population. Second, this retrospective study of patients who underwent RFCA with 2 types of catheters (CF vs conventional) was neither randomized nor controlled and involved no matching between the 2 groups. A large-scale, prospective, randomized, controlled study to confirm our present results is warranted. Third, our study was designed as a pilot study to investigate the influence of CF measurement on acute procedural parameters in patients undergoing RFCA of PVCs initially. Monitoring of the long-term clinical outcomes in these patients is warranted. Finally, despite the involvement of experienced electrophysiologists, procedure, fluoroscopy, and ablation times might have been affected by other operator-related factors such as length of experience in ventricular ablation, familiarity with the novel CF-sensing catheter, and variation in anatomic access.

Conclusion

Ablation of PVCs originating from the RVOT with an irrigated CF-sensing catheter that enables measurement of CF and calculation of FTI in real time can shorten overall procedure, fluoroscopy, and ablation times without increasing the risk of recurrence or complications. Prospective, randomized studies in larger populations are warranted.

References

References
1.
Haines
DE.
Determinants of lesion size during radiofrequency catheter ablation: the role of electrode-tissue contact pressure and duration of energy delivery
.
J Cardiovasc Electrophysiol
1991
;
2
(
6
):
509
15
.
2.
Barnett
AS
,
Bahnson
TD
,
Piccini
JP.
Recent advances in lesion formation for catheter ablation of atrial fibrillation
.
Circ Arrhythm Electrophysiol
2016
;
9
(
5
):
e003299
.
3.
Yokoyama
K
,
Nakagawa
H
,
Shah
DC
,
Lambert
H
,
Leo
G
,
Aeby
N
,
et al
.
Novel contact force sensor incorporated in irrigated radiofrequency ablation catheter predicts lesion size and incidence of steam pop and thrombus
.
Circ Arrhythm Electrophysiol
2008
;
1
(
5
):
354
62
.
4.
Shah
DC
,
Lambert
H
,
Nakagawa
H
,
Langenkamp
A
,
Aeby
N
,
Leo
G.
Area under the real-time contact force curve (force-time integral) predicts radiofrequency lesion size in an in vitro contractile model
.
J Cardiovasc Electrophysiol
2010
;
21
(
9
):
1038
43
.
5.
Thiagalingam
A
,
D'Avila
A
,
Foley
L
,
Guerrero
JL
,
Lambert
H
,
Leo
G
,
et al
.
Importance of catheter contact force during irrigated radiofrequency ablation: evaluation in a porcine ex vivo model using a force-sensing catheter
.
J Cardiovasc Electrophysiol
2010
;
21
(
7
):
806
11
.
6.
Seiler
J
,
Roberts-Thomson
KC
,
Raymond
JM
,
Vest
J
,
Delacretaz
E
,
Stevenson
WG.
Steam pops during irrigated radiofrequency ablation: feasibility of impedance monitoring for prevention
.
Heart Rhythm
2008
;
5
(
10
):
1411
6
.
7.
Reichlin
T
,
Knecht
S
,
Lane
C
,
Kuhne
M
,
Nof
E
,
Chopra
N
,
et al
.
Initial impedance decrease as an indicator of good catheter contact: insights from radiofrequency ablation with force sensing catheters
.
Heart Rhythm
2014
;
11
(
2
):
194
201
.
8.
Kuck
KH
,
Reddy
VY
,
Schmidt
B
,
Natale
A
,
Neuzil
P
,
Saoudi
N
,
et al
.
A novel radiofrequency ablation catheter using contact force sensing: Toccata study
.
Heart Rhythm
2012
;
9
(
1
):
18
23
.
9.
Sarkozy
A
,
Shah
D
,
Saenen
J
,
Sieira
J
,
Phlips
T
,
Boris
W
,
et al
.
Contact force in atrial fibrillation: role of atrial rhythm and ventricular contractions: Co-Force Atrial Fibrillation Study
.
Circ Arrhythm Electrophysiol
2015
;
8
(
6
):
1342
50
.
10.
Lin
H
,
Chen
YH
,
Hou
JW
,
Lu
ZY
,
Xiang
Y
,
Li
YG.
Role of contact force-guided radiofrequency catheter ablation for treatment of atrial fibrillation: a systematic review and meta-analysis
.
J Cardiovasc Electrophysiol
2017
;
28
(
9
):
994
1005
.
11.
Valk
SD
,
de Groot
NM
,
Jordaens
L.
Catheter ablation of right ventricular outflow tract tachycardia using contact force guidance
.
Neth Heart J
2014
;
22
(
10
):
460
2
.
12.
Akca
F
,
Janse
P
,
Theuns
DA
,
Szili-Torok
T.
A prospective study on safety of catheter ablation procedures: contact force guided ablation could reduce the risk of cardiac perforation
.
Int J Cardiol
2015
;
179
:
441
8
.
13.
Liang
JJ
,
Santangeli
P.
Contact force sensing during atrial fibrillation ablation: clinical experience and effects on outcomes
.
Expert Rev Cardiovasc Ther
2016
;
14
(
6
):
749
59
.
14.
Weiss
JP
,
May
HT
,
Bair
TL
,
Crandall
BG
,
Cutler
MJ
,
Day
JD
,
et al
.
A comparison of remote magnetic irrigated tip ablation versus manual catheter irrigated tip catheter ablation with and without force sensing feedback
.
J Cardiovasc Electrophysiol
2016
;
27
Suppl 1
:
S5
S10
.
15.
Stabile
G
,
Di Donna
P
,
Schillaci
V
,
Di Monaco
A
,
Iuliano
A
,
Caponi
D
,
et al
.
Safety and efficacy of pulmonary vein isolation using a surround flow catheter with contact force measurement capabilities: a multicenter registry
.
J Cardiovasc Electrophysiol
2017
;
28
(
7
):
762
7
.
16.
Reddy
VY
,
Shah
D
,
Kautzner
J
,
Schmidt
B
,
Saoudi
N
,
Herrera
C
,
et al
.
The relationship between contact force and clinical outcome during radiofrequency catheter ablation of atrial fibrillation in the TOCCATA study
.
Heart Rhythm
2012
;
9
(
11
):
1789
95
.
17.
Sanchez
JE
,
Kay
GN
,
Benser
ME
,
Hall
JA
,
Walcott
GP
,
Smith
WM
,
Ideker
RE.
Identification of transmural necrosis along a linear catheter ablation lesion during atrial fibrillation and sinus rhythm
.
J Interv Card Electrophysiol
2003
;
8
(
1
):
9
17
.
18.
Reithmann
C
,
Remp
T
,
Hoffmann
E
,
Matis
T
,
Wakili
R
,
Steinbeck
G.
Different patterns of the fall of impedance as the result of heating during ostial pulmonary vein ablation: implications for power titration
.
Pacing Clin Electrophysiol
2005
;
28
(
12
):
1282
91
.
19.
Otomo
K
,
Uno
K
,
Fujiwara
H
,
Isobe
M
,
Iesaka
Y.
Local unipolar and bipolar electrogram criteria for evaluating the transmurality of atrial ablation lesions at different catheter orientations relative to the endocardial surface
.
Heart Rhythm
2010
;
7
(
9
):
1291
300
.
20.
Ullah
W
,
Hunter
RJ
,
Baker
V
,
Dhinoja
MB
,
Sporton
S
,
Earley
MJ
,
Schilling
RJ.
Target indices for clinical ablation in atrial fibrillation: insights from contact force, electrogram, and biophysical parameter analysis
.
Circ Arrhythm Electrophysiol
2014
;
7
(
1
):
63
8
.
21.
Sulkin
MS
,
Laughner
JI
,
Hilbert
S
,
Kapa
S
,
Kosiuk
J
,
Younan
P
,
et al
.
Novel measure of local impedance predicts catheter-tissue contact and lesion formation
.
Circ Arrhythm Electrophysiol
2018
;
11
(
4
):
e005831
.
22.
Wakili
R
,
Clauss
S
,
Schmidt
V
,
Ulbrich
M
,
Hahnefeld
A
,
Schussler
F
,
et al
.
Impact of real-time contact force and impedance measurement in pulmonary vein isolation procedures for treatment of atrial fibrillation
.
Clin Res Cardiol
2014
;
103
(
2
):
97
106
.

Author notes

From: Cardiovascular Department (Drs. Chang, Gao, Liu, Xia, Xiao, Yang, Yin, and Zhao), The First Affiliated Hospital of Dalian Medical University, 166011 Dalian, People's Republic of China; and Basic Electrophysiological Research Laboratory (Drs. Chen, Cheng, and Xi), Texas Heart Institute, Houston, Texas 77030

Drs. Zhao and Liu contributed equally to this work.

Presented in part at the American Heart Association Scientific Sessions and Resuscitation Science Symposium, New Orleans, 12–14 November 2016.

This work was supported by grants from Liaoning Province (201201009, L2011154, 2015020259, 201501853, L2016013), the Chinese Ministry of Health (w201003), and Dalian Medical University.