**4. Morphologic and histogenetic studies of experimental NPC**

In the normal rat, cephalic and middle portion of nasopharyngeal cavity is covered by ciliated columnar epithelium, while caudal portion is covered by squamous and columnar epithelium.

#### **4.1 Pathology of the experimental NPC**

68 Carcinogenesis, Diagnosis, and Molecular Targeted Treatment for Nasopharyngeal Carcinoma

pattern observed in this experiment was somewhat different from that of NPC induced by diethylnitrosamine. The cancer cells invaded the stroma in trabecular or branched cords, but more frequently grew intraepithelially to form patches of masses before invading the stroma. A lymphatic cancer embolus consisting of poorly differentiated cancer cells was found in a case of invasive NPC. No metastatic foci could be found in the lymph nodes of

Fig. 1. DNP- induced early nasopharyngeal cancer **(**Original magnification, ×200**).** Wistar rats were injected subcutaneously with DNP in a stumped needle at a dosage of 40 mg/kg, twice a week and 38 times in total to accumulation dose of 99.5–122 mg per rat. The rats were sacrificed to collect nasopharyngeal samples at 308 days after DNP injections. Nasopharyngeal samples were histopathologically examined under microscope 200.

Most of nasal cavity tumors developed from nasal turbinates. Occasionally, tumors destroyed nasal bone and bulged out as a local prominence. The soft tumor mass with a pink-grey tint frequently destroyed cribriform plate and invaded or replaced olfactory bulb, but it was well demarcated from brain tissue. The tumors of nasal cavity were of different histological types, including squamous cell carcinoma, adenocarcinoma, and the so-called "olfactory neuroepithelial tumor." The olfactory neuroepithelial tumors had a pleomorphic histology that suggested that these tumors originated from epithelia. On gross examination, most of esophageal tumors were multiple papillomas, that developed mainly from the upper and middle parts of esophagus; these papillomas were histologically confirmed as

Some remaining points need to be clarified: (a) the induced rat NPC was of a squamous cell type. Most of tumors were well-differentiated and developed from base and lateral side of nasopharynx, and some co-existed with squamous carcinoma of soft palate mucosa. There was rarely distant metastasis; in this regard NPC induced by nitroso-compounds was somewhat different from human NPC. (b) DNP-induced NPC and cancer of nasal cavities and esophagus in rats resulted from a subcutanrouic effect. (c) Nitroso-compounds often exist as cis- and trans-isomers, and rapid axial-equatorial conversion of dinitrospiperazin

and the respective carcinogenicity of these isomers should be explored in the future.

this animal [72, 73].

squamous cell carcinoma [73].

In the soft palates of tumor-bearing rats, nodules ranging from1-2 mm to 1cm in diameter were observed by palpation. Some of the invasive tumors were ulcerative cauliflowerlike or nodular in appearance when the soft palate was cut open, cancerous ulcer with elevated irregular edge was observed when cancerous ulcer with elevated irregular edge was cut open [24]. Under the microscope, cancer cell foci in different stages of differentiation could be found in many serial sections, indicating the multiple growth pattern of malignancy (Fig.1. DNP- induced early nasopharyngeal cancer. Wistar rats were injected subcutaneously with DNP in a stumped needle at a dosage of 40 mg/kg, twice a week and 38 times in total to accumulation dose of 99.5–122 mg per rat. The rats were sacrificed to collect nasopharyngeal samples at 308 days after DNP injections. Nasopharyngeal samples were histopathologically examined under microscope 200). A localized solitary malignancy was observed in rats that received carcinogen for a short period. Tumors often developed at the base of nasopharynx or at the junction of the base and lateral side of nasopharyngeal wall. Carcinomas in situ, and early invasive or infiltrative carcinomas were observed, and were of a well-differentiated squamous cell type [73].

The diagnostic criterion of experimental NPC is generally that of human nasopharyngeal cancer, but the structural characteristics of rat nasopharynx must also be taken into consideration. For example, the epithelium at nasopharynx base, especially at nasopharynx roof next to skull base, consists of two or three layers of cells surrounded by bony plates[59]. The observed malignancy may be composed of basal cells with thin solid strands of cancer cells.

Carcinoma in situ, early invasive carcinoma, and infiltrative carcinoma were observed in tumor-bearing rats. Carcinoma in situ might arise from epithelium of normal thickness, or from highly hyperplastic or atrophic epithelium (Fig.2. DEN- induced nasopharyngeal cancer. Wistar rats were injected subcutaneously with DEN in a stumped needle at a dosage of 40 mg/kg, twice a week and 38 times in total to accumulation dose of 89.8–119.3 mg per rat. The rats were sacrificed to collect nasopharyngeal samples at 365 days after DEN injections. Nasopharyngeal samples were histopathologically examined under microscope). It often originates from columnar epithelium and evolves through squamous cell metaplasia, atypical hyperplasia, and Grade I and II anaplasia. The early invasive lesions were of two types: arising from basal layer with normal looking superficial layers, or extending into stroma (primarily seen in the DEN group). Downward growth of the lesion, which formed trabecular, branching, rounded, or small square nests, infiltrated into the stroma. The neoplasms were usually moderately or even poorly differentiated squamous carcinomas. This downward growth of neoplasms, which often occurs in human NPC, may cause difficulty in early detection or in producing a good smear of exfoliative cells for cytological diagnosis [57,72,73].

Chemical Carcinogenesis and Nasopharyngeal Carcinoma 71

arising from different portions of the epithelium, might finally fuse together to form a large solid mass of tumor [57, 72, 73]. It was evident that most of the lesions were multicentric rather than unicentric in origin. Sometimes the same cancer cell focus contained not only undifferentiated fusiform cells and small cornified epithelial pearls, but also small cysts containing mucinous material. These kinds of lesions may be seen in human NPC and are

The well-known two-stage concept suggested by Berenblum et al [60] was developed from experimental study of carcinogenesis. They proposed that carcinogenesis might be divided into two different but related stages, a stage of specific initiation and a stage of relatively non-specific promotion. However, on the basis of epidemiological data and statistical studies of human cancer, carcinogenesis is generally considered a multi-hit/multi-step process. The multistage theory of carcinogenesis proposed by Amitage and Dell is representative of this concept [61]. Therefore, it is interesting to analyze the carcinogenesis

Yao et al [73] reported the results of statistical analysis of experimental NPC induced by DNP. The corrected cumulative percentage of dead rats with NPC was calculated and fitted with Weibull and lognormal distributions. The data fitted both distributions well as verified by 2-test, but the value of 2 was smaller in the Wieibull distribution than in the lognormal distribution. The mathematical expression of NPC carcinogenesis in rats according to

G=1-e -2.55×10-4 (t-192) 1.69

Peto [62] and Emmelot et al[63] suggested that Weibull's distribution represents the number of "hits" or "stages" in cancer development. Therefore, the above mathematical expression seems to indicate that rat NPC development experienced two hits. Using retrospective survey data of cancer mortalities in Hunan Province, Yao [73] analyzed the age distribution of NPC mortality and proposed that the development of NPC needs three hits. Hence, the relationship and differences between carcinogenesis in experimental NPC and human NPC

**4.4.1 Atypical cytokinetics of nasopharyngeal epithelium in rats treated with DNP** 

Cytokinetic studies may be helpful in elucidating the mechanism of carcinogenesis and providing important data for improving chemotherapeutic regimens for malignancies. In 1981, Chen et al [74] studied the cytokinetics of nasopharyngeal epithelium of rats treated with DNP using stathmokinetic and autoradiographic techniques. The mitotic blocking agent selected was vincristine, which had an optimal dose of 0.83 mg/kg body weight and a blocking effect that lasted 10 to 12 h. The metaphase chromosomes of mitotic cells appeared as deeply stained rosette-like structures under the microscope. The cytokinetic parameters of normal and DNP\_treated rats were measured *in vivo* using 3H\_TdR labeling in

G, corrected cumulative percentage of the dead rats bearing NPC; t, time in days

considered to represent the biphasic differentiation of nasopharyngeal epithelium.

**4.3 Statistical analysis of NPC carcinogenesis** 

of experimental NPC by means of mathematical statistics.

**4.4 Carcinogenesis mechanism of experimental NPC** 

Weibull distribution was as follows:

need to be explored further.

Fig. 2**.** DEN- induced nasopharyngeal cancer **(**Original magnification, ×400**).** Wistar rats were injected subcutaneously with DEN in a stumped needle at a dosage of 40 mg/kg, twice a week and 38 times in total to accumulation dose of 89.8–119.3 mg per rat. The rats were sacrificed to collect nasopharyngeal samples at 365 days after DEN injections. Nasopharyngeal samples were histopathologically examined under microscope 400.

Arising from a dyaplasitc squamous epithelium, cancer protruded into nasopharyngeal cavity at the one hand and invaded the stroma at the other. This growth pattern occurred more frequently in the DNP group, and sometimes formed papillary carcinomas [57,72,73]. The infiltrative carcinomas were usually of a squamous cell type, although basal cell carcinoma was observed occasionally. This kind of tumor growth could be seen in many serial sections. The cancer cells forming trabecular, solid masses, or patches were not well differentiated at the periphery but had cornified cells in the central portion. Cancer cells invaded salivary glands, muscle, and nerve bundles with occasional lymphatic emboli but without widespread metastases to the lymph nodes. The primary site or origin of the infiltrative tumor was usually difficult to recognize, but could occasionally be traced by serial sectioning.

#### **4.2 The histogenesis of experimental NPC**

In early lesions, dysplasia occurred in the ciliated columnar, transitional, or squamous epithelium, and was possibly preceded or accompanied by squamous metaplasia. 3Hthymidine autoradiography showed that a single dose of dinitrospiperazine could cause DNA damage of the squamous or transitional cells followed by unscheduled DNA synthesis [74]. This may give some clues to NPC carcinogenesis. Hyperplastic, dysplastic, and neoplastic foci often co-existed without sharp demarcation. The cancer cell population was more localized and clearly demarcated from the surrounding normal epithelium or stromal tissues. Inflammatory infiltration into the surrounding normal stroma was rarely found except in ulcerative carcinomas.

NPC was present in multiple serial sections due to the extension of a large tumor mass or multiple diffusely scattered lesions. Some of the irregular cords or trabeculae of tumor cells,

Fig. 2**.** DEN- induced nasopharyngeal cancer **(**Original magnification, ×400**).** Wistar rats were injected subcutaneously with DEN in a stumped needle at a dosage of 40 mg/kg, twice a week and 38 times in total to accumulation dose of 89.8–119.3 mg per rat. The rats were

Arising from a dyaplasitc squamous epithelium, cancer protruded into nasopharyngeal cavity at the one hand and invaded the stroma at the other. This growth pattern occurred more frequently in the DNP group, and sometimes formed papillary carcinomas [57,72,73]. The infiltrative carcinomas were usually of a squamous cell type, although basal cell carcinoma was observed occasionally. This kind of tumor growth could be seen in many serial sections. The cancer cells forming trabecular, solid masses, or patches were not well differentiated at the periphery but had cornified cells in the central portion. Cancer cells invaded salivary glands, muscle, and nerve bundles with occasional lymphatic emboli but without widespread metastases to the lymph nodes. The primary site or origin of the infiltrative tumor was usually difficult to recognize, but could occasionally be traced by

In early lesions, dysplasia occurred in the ciliated columnar, transitional, or squamous epithelium, and was possibly preceded or accompanied by squamous metaplasia. 3Hthymidine autoradiography showed that a single dose of dinitrospiperazine could cause DNA damage of the squamous or transitional cells followed by unscheduled DNA synthesis [74]. This may give some clues to NPC carcinogenesis. Hyperplastic, dysplastic, and neoplastic foci often co-existed without sharp demarcation. The cancer cell population was more localized and clearly demarcated from the surrounding normal epithelium or stromal tissues. Inflammatory infiltration into the surrounding normal stroma was rarely found

NPC was present in multiple serial sections due to the extension of a large tumor mass or multiple diffusely scattered lesions. Some of the irregular cords or trabeculae of tumor cells,

sacrificed to collect nasopharyngeal samples at 365 days after DEN injections. Nasopharyngeal samples were histopathologically examined under microscope 400.

serial sectioning.

**4.2 The histogenesis of experimental NPC** 

except in ulcerative carcinomas.

arising from different portions of the epithelium, might finally fuse together to form a large solid mass of tumor [57, 72, 73]. It was evident that most of the lesions were multicentric rather than unicentric in origin. Sometimes the same cancer cell focus contained not only undifferentiated fusiform cells and small cornified epithelial pearls, but also small cysts containing mucinous material. These kinds of lesions may be seen in human NPC and are considered to represent the biphasic differentiation of nasopharyngeal epithelium.

#### **4.3 Statistical analysis of NPC carcinogenesis**

The well-known two-stage concept suggested by Berenblum et al [60] was developed from experimental study of carcinogenesis. They proposed that carcinogenesis might be divided into two different but related stages, a stage of specific initiation and a stage of relatively non-specific promotion. However, on the basis of epidemiological data and statistical studies of human cancer, carcinogenesis is generally considered a multi-hit/multi-step process. The multistage theory of carcinogenesis proposed by Amitage and Dell is representative of this concept [61]. Therefore, it is interesting to analyze the carcinogenesis of experimental NPC by means of mathematical statistics.

Yao et al [73] reported the results of statistical analysis of experimental NPC induced by DNP. The corrected cumulative percentage of dead rats with NPC was calculated and fitted with Weibull and lognormal distributions. The data fitted both distributions well as verified by 2-test, but the value of 2 was smaller in the Wieibull distribution than in the lognormal distribution. The mathematical expression of NPC carcinogenesis in rats according to Weibull distribution was as follows:

$$\text{G=1-e } \text{-2.55} \times 10^4 \text{ (t-192) } ^{1.69}\text{ }$$

G, corrected cumulative percentage of the dead rats bearing NPC; t, time in days

Peto [62] and Emmelot et al[63] suggested that Weibull's distribution represents the number of "hits" or "stages" in cancer development. Therefore, the above mathematical expression seems to indicate that rat NPC development experienced two hits. Using retrospective survey data of cancer mortalities in Hunan Province, Yao [73] analyzed the age distribution of NPC mortality and proposed that the development of NPC needs three hits. Hence, the relationship and differences between carcinogenesis in experimental NPC and human NPC need to be explored further.

#### **4.4 Carcinogenesis mechanism of experimental NPC**

#### **4.4.1 Atypical cytokinetics of nasopharyngeal epithelium in rats treated with DNP**

Cytokinetic studies may be helpful in elucidating the mechanism of carcinogenesis and providing important data for improving chemotherapeutic regimens for malignancies. In 1981, Chen et al [74] studied the cytokinetics of nasopharyngeal epithelium of rats treated with DNP using stathmokinetic and autoradiographic techniques. The mitotic blocking agent selected was vincristine, which had an optimal dose of 0.83 mg/kg body weight and a blocking effect that lasted 10 to 12 h. The metaphase chromosomes of mitotic cells appeared as deeply stained rosette-like structures under the microscope. The cytokinetic parameters of normal and DNP\_treated rats were measured *in vivo* using 3H\_TdR labeling in

Chemical Carcinogenesis and Nasopharyngeal Carcinoma 73

Carcinogenesis is a multistep process. Cytokinetic parameters may be changed with morphological progression from normal epithelium, hyperplasia, atypical hyperplasia, to carcinoma. Atypical cytokinetics, increased LI, expansion of the proliferation compartment, and prolongation of S phase were observed in nasopharyngeal epithelium after DNP

The adult rats were sacrificed at 4 hrs after a single injection of DNP intramuscularly, and then the tissue fragments of nasopharynx, esophayus, kidney and liver were cultured in 3H \_ thymidine \_ 199 medium for 10 hrs. Autoradiography was performed to examine 3H \_ thymidine labeled epithelial cells. The experimental data indicated that LI of nasopharyngeal epithelium markedly decreased to 2.40% in the experimental group with DNP treatment compared with the saline control, which suggested that a single dose of DNP inhibits DNA synthesis of nasopharyngeal epithelium. This inhibition may be result of DNA damage, providing a key link in carcinogenesis process of experimental

Le et al [75] injected 0.25% DNP solution (15 mg/kg bodyweight) into the dorsum of rats. Animals were sacrificed by cervical dislocation at 4 h, 79 h, and 124 h after injection. 3H\_TdR (0.5 mCi/kg bodyweight) was injected intraperitoneally 1 h before the animals were sacrificed. The nasopharynx, esophagus, and forestomach were sectioned and processed histologically and autoradiographically. A count of 1,000 consecutive basal cells was performed in the epithelia of nasopharynx, esophagus, and fore stomach of each animal. The labeled cells (>5 silver grains/nucleus) were scored and the LIs were expressed as percentages of the mean value for the controls of the same experiment. LI of squamous epithelia in the base of nasopharynx and esophagus significantly decreased 4 h after DNP injection and recovered gradually by the 3rd and 5th days; the LI of squamous epithelium in nasopharynx base was actually significantly higher than that of the control at 5 days after DNP injection. LI in the squamous epithelium of forestomach did not significantly change at

UDS was detected autoradiographically. 30, 50 or 80 mg/kg in 0.5% DNP was injected subcutaneously into the dorsum of rats [64] and an equivalent amount of saline was injected into rats of the control group. The rats were sacrificed 2 h after injection and small epithelial tissue fragments of the nasal concha, soft palate, esophagus, and forestomach, as well as the basal and lateral side of nasopharynx, were removed immediately. The tissues were processed histologically and autoradiographically. It is easy to identify the nuclei of S phase cells by their extremely heavy labeling. UDS was considered present if there were at least nuclei, each covered by 5-20 silver grains, in a cluster and the number of grains covering the nuclei was no more than that of the background (Fig.4. UDS autoradiograph of nasopharyngeal squamous epithelium**.** DNP was injected subcutaneously into the dorsum

**thymidine incorporation into DNA of rat nasopharyngeal** 

treatment.

**epithelium** 

**4.4.2 Effect of DNP on <sup>3</sup>**

nasopharyngeal cancer [73].

**H\_**

**4.4.3 Organotropic action of nitroso-compound carcinogensis** 

4 h or on the 3rd day, but declined significantly 5 days after DNP treatment.

**4.4.4 Unscheduled DNA Synthesis (UDS)** 

combination with stathmokinetics. After DNP treatment there were many hyperplastic foci in nasopharyngeal squamous epithelium. Within the hyperplastic foci and apparently normal nasopharyngeal epithelium there was a significant increase in the number of labeled basal cells, but no change in labeling index (LI) of the transitional epithelium lining the lateral side of nasopharynx.

In normal rats, LI was highest in the squamous epithelium lining nasopharynx lateral side, followed by the base and then the transitional epithelium. The differences between these were highly significant. After DNP treatment, LI of both the lateral and base side increased, and there was no difference between them. This suggested that LI of the bottom side increased more significantly than that of the lateral side. In the normal rats, 3H-TdR labeled cells in nasopharyngeal squamous epithelium were confined to the basal layer (Fig.3. Autoradiaograph of nasopharyngeal squamous epithelium of normal rat, showing labeled cells in basal cell layer**.** Rats were treated with vincristine at 0.83 mg/kg body weight for 12 h, and then 3H\_TdR was injected for 1 h. The rats were sacrificed to collect nasopharyngeal samples. The nasopharyngeal samples were histopathologically examined. Cells with 3H\_TdR labeling were observed under microscope.). After DNP treatment, there was a significant increase in the number of labeled basal cells. Moreover, a few cells in the prickle cell layer were found to be labeled, indicating that the proliferation compartment had expanded. This phenomenon suggested that stem cells in the G0 stage might enter the proliferation stage and the two daughter cells of mitosis had proliferative ability. S phase of nasopharyngeal epithelium located at the lateral bottom side was prolonged from 6.7 to 9 h after DNP treatment. The time of cell cycle can be calculated according to the formula if the proliferation of nasopharyngeal epithelium in normal and DNP-treated rats remains in steady state.

Fig. 3. Autoradiaograph of nasopharyngeal squamous epithelium of normal rat, showing labeled cells in basal cell layer **(**Original magnification, ×400**).** For mitotic blocking of nasopharyngeal epithelium, rats were treated with vincristine at 0.83 mg/kg body weight for 12 h. 3H\_TdR (0.5 mCi/kg bodyweight) was intraperitoneally injected into the rats for 1 h. The rats were sacrificed to collect nasopharyngeal samples. The nasopharyngeal samples were histopathologically examined. Cells with 3H\_TdR labeling were observed under microscope 400

combination with stathmokinetics. After DNP treatment there were many hyperplastic foci in nasopharyngeal squamous epithelium. Within the hyperplastic foci and apparently normal nasopharyngeal epithelium there was a significant increase in the number of labeled basal cells, but no change in labeling index (LI) of the transitional epithelium lining the

In normal rats, LI was highest in the squamous epithelium lining nasopharynx lateral side, followed by the base and then the transitional epithelium. The differences between these were highly significant. After DNP treatment, LI of both the lateral and base side increased, and there was no difference between them. This suggested that LI of the bottom side increased more significantly than that of the lateral side. In the normal rats, 3H-TdR labeled cells in nasopharyngeal squamous epithelium were confined to the basal layer (Fig.3. Autoradiaograph of nasopharyngeal squamous epithelium of normal rat, showing labeled cells in basal cell layer**.** Rats were treated with vincristine at 0.83 mg/kg body weight for 12 h, and then 3H\_TdR was injected for 1 h. The rats were sacrificed to collect nasopharyngeal samples. The nasopharyngeal samples were histopathologically examined. Cells with 3H\_TdR labeling were observed under microscope.). After DNP treatment, there was a significant increase in the number of labeled basal cells. Moreover, a few cells in the prickle cell layer were found to be labeled, indicating that the proliferation compartment had expanded. This phenomenon suggested that stem cells in the G0 stage might enter the proliferation stage and the two daughter cells of mitosis had proliferative ability. S phase of nasopharyngeal epithelium located at the lateral bottom side was prolonged from 6.7 to 9 h after DNP treatment. The time of cell cycle can be calculated according to the formula if the proliferation

of nasopharyngeal epithelium in normal and DNP-treated rats remains in steady state.

Fig. 3. Autoradiaograph of nasopharyngeal squamous epithelium of normal rat, showing labeled cells in basal cell layer **(**Original magnification, ×400**).** For mitotic blocking of nasopharyngeal epithelium, rats were treated with vincristine at 0.83 mg/kg body weight for 12 h. 3H\_TdR (0.5 mCi/kg bodyweight) was intraperitoneally injected into the rats for 1 h. The rats were sacrificed to collect nasopharyngeal samples. The nasopharyngeal samples were histopathologically examined. Cells with 3H\_TdR labeling were observed

lateral side of nasopharynx.

under microscope 400

Carcinogenesis is a multistep process. Cytokinetic parameters may be changed with morphological progression from normal epithelium, hyperplasia, atypical hyperplasia, to carcinoma. Atypical cytokinetics, increased LI, expansion of the proliferation compartment, and prolongation of S phase were observed in nasopharyngeal epithelium after DNP treatment.

#### **4.4.2 Effect of DNP on <sup>3</sup> H\_ thymidine incorporation into DNA of rat nasopharyngeal epithelium**

The adult rats were sacrificed at 4 hrs after a single injection of DNP intramuscularly, and then the tissue fragments of nasopharynx, esophayus, kidney and liver were cultured in 3H \_ thymidine \_ 199 medium for 10 hrs. Autoradiography was performed to examine 3H \_ thymidine labeled epithelial cells. The experimental data indicated that LI of nasopharyngeal epithelium markedly decreased to 2.40% in the experimental group with DNP treatment compared with the saline control, which suggested that a single dose of DNP inhibits DNA synthesis of nasopharyngeal epithelium. This inhibition may be result of DNA damage, providing a key link in carcinogenesis process of experimental nasopharyngeal cancer [73].

#### **4.4.3 Organotropic action of nitroso-compound carcinogensis**

Le et al [75] injected 0.25% DNP solution (15 mg/kg bodyweight) into the dorsum of rats. Animals were sacrificed by cervical dislocation at 4 h, 79 h, and 124 h after injection. 3H\_TdR (0.5 mCi/kg bodyweight) was injected intraperitoneally 1 h before the animals were sacrificed. The nasopharynx, esophagus, and forestomach were sectioned and processed histologically and autoradiographically. A count of 1,000 consecutive basal cells was performed in the epithelia of nasopharynx, esophagus, and fore stomach of each animal. The labeled cells (>5 silver grains/nucleus) were scored and the LIs were expressed as percentages of the mean value for the controls of the same experiment. LI of squamous epithelia in the base of nasopharynx and esophagus significantly decreased 4 h after DNP injection and recovered gradually by the 3rd and 5th days; the LI of squamous epithelium in nasopharynx base was actually significantly higher than that of the control at 5 days after DNP injection. LI in the squamous epithelium of forestomach did not significantly change at 4 h or on the 3rd day, but declined significantly 5 days after DNP treatment.
